YEARS OF EUROPEAN ONLINE ANNÉES DE EN LIGNE ...

Speeches and proceedings
25th anniversary of EU law online
23 and 24 November 2006
Luxembourg
25
YEARS OF
EUROPEAN
ONLINE
THE EVENT
25
ANNÉES DE
D RO I T
EUROPÉEN
EN LIGNE
L’ÉVÉNEMENT
Discours et actes
25 e anniversaire du droit européen en ligne
23 et 24 novembre 2006
Luxembourg
01_2007_5222_txt_ML.indd 1 6-12-2007 15:12:38

the views expressed only bind the author(s) and should not be considered as constituting
the oficial position of the European Commission.
Les avis exprimés n’engagent que l’auteur (les auteurs) et ne sauraient être considérés
comme constituant une prise de position oficielle de la Commission européenne.
A great deal of additional information on the European Union is available on the Internet.
It can be accessed through the Europa server (http://europa.eu).
Cataloguing data can be found at the end of this publication.
Luxembourg: Ofice for Oficial Publications of the European Communities, 2007
© European Communities, 2007
Reproduction is authorised provided the source is acknowledged.
De nombreuses autres informations sur l’Union européenne sont disponibles sur l’internet
via le serveur Europa (http://europa.eu).
Une iche bibliographique igure à la in de l’ouvrage.
Luxembourg: Ofice des publications oficielles des Communautés européennes, 2007
© Communautés européennes, 2007
Reproduction autorisée, moyennant mention de la source
ISBN 978-92-78-40413-0
Printed in Italy
PRINtED ON whItE ChLORINE-fREE PAPER
ImPRImé SUR PAPIER BLANChI SANS ChLORE
01_2007_5222_txt_ML.indd 2 6-12-2007 15:12:38

25
YEARS OF
ONLINE
THE EVENT
25
ANNÉES DE
DROIT EUROPÉEN
EN LIGNE
L’ÉVÉNEMENT
OffICE fOR OffICIAL PUBLICAtIONS
Of thE EUROPEAN COmmUNItIES
OffICE DES PUBLICAtIONS OffICIELLES
DES COmmUNAUtéS EUROPéENNES
Luxembourg 2007
01_2007_5222_txt_ML.indd 3 6-12-2007 15:12:38

:
Welcome of His Royal Highness
the Grand-Duke of Luxembourg
by Thomas L. Cranfield,
Director of the Publications Office
01_2007_5222_txt_ML.indd 4 6-12-2007 15:12:45

INTRODUCTION
APRèS LA PUBLICAtION d’une première brochure le jour où l’Ofice des publications
célébrait le 25 e anniversaire du droit européen en ligne, il a semblé opportun de publier
une seconde brochure reprenant l’ensemble des discours et interventions prononcés à
cette occasion, les 23 et 24 novembre 2006.
Le 25 e anniversaire du droit européen en ligne a été marqué par une séance académique
le 23 novembre 2006, à l’hémicycle du Kirchberg, à Luxembourg, honorée de la
présence de Son Altesse Royale le Grand-Duc de Luxembourg. La séance académique
était placée sous la présidence de m me Kirsti Rissanen, secrétaire permanente du ministère
de la justice de finlande. Les quelque 400 personnes de l’assistance ont pu écouter
successivement m. thomas L. Cranield, directeur de l’Ofice des publications, m me
Claire-françoise Durand, directeur général adjoint au service juridique de la Commission,
qui a pris la parole au nom de m. José manuel Barroso, président de la Commission,
m me hélène Bernet, directeur honoraire de la Commission, qui est à l’origine du
traitement électronique du droit européen dans les années 60 et 70, m. maximilian herberger,
professeur de droit à l’université de la Sarre, titulaire de la chaire de droit civil, de
théorie du droit et d’informatique juridique, et enin m me Kirsti Rissanen.
Le groupe de travail «Informatique juridique» du Conseil, qui tenait exceptionnellement
sa réunion semestrielle à Luxembourg, a consacré la matinée du 24 novembre à
EUR-Lex et à l’accès au droit. La réunion a été suivie l’après-midi d’un séminaire.
L’Ofice des publications remercie encore ici tous les orateurs, qui ont en outre accepté
que leurs contributions soient publiées.
fOLLOwING PUBLICAtION of a irst brochure on the day the Publications Ofice celebrated
the 25th anniversary of EU law online, it seemed appropriate to publish a second
brochure containing all speeches and presentations of the event, which took place on 23
and 24 November 2006.
the 25th anniversary of EU law online was celebrated in an academic session at the
Kirchberg hemicycle, in Luxembourg, which his Royal highness Grand-Duke of Luxembourg
honoured with his presence. the academic session was chaired by ms Kirsti
Rissanen, Permanent Secretary of the ministry of Justice of finland. the audience of
some 400 people could successively listen to mr thomas L. Cranield, Director of the
Publications Ofice, ms Claire-françoise Durand, Deputy Director-General in the Legal
Service of the Commission, who spoke in the name of mr José manuel Barroso, President
of the Commission, ms hélène Bernet, honorary Director of the Commission,
who took the irst steps towards the electronic processing of European law in the 1960s
and 1970s and mr maximilian herberger, professor of private law, law theory and legal
informatics at the University of Saarland.
the Council working Party on Legal Data Processing, which held its half-yearly
meeting in Luxembourg for the occasion, dedicated the morning of 24 November to EUR-
Lex and to the access to law. the meeting was followed in the afternoon by a workshop.
the Publications Ofice once again thanks all speakers for their participation and
for agreeing to the publication of their contributions.
01_2007_5222_txt_ML.indd 5 6-12-2007 15:12:45

25
YEARS OF
EUROPEAN
ONLINE
1. His Royal Highness the Grand-Duke of Luxembourg,
Kirsti Rissanen and Thomas L. Cranfield
1
3 4
5
2
2. Hemicycle
3. Hemicycle
4. His Royal Highness the Grand-Duke of Luxembourg
and Thomas L. Cranfield
5. Maximilian Herberger, Kirsti Rissanen and Hélène Bernet
01_2007_5222_txt_ML.indd 6 6-12-2007 15:13:02

wORKShOP
Legal XmL — Use of XmL for the production
and distribution of oficial gazettes
Søren Broberg Nielsen ............................................................................................. 77
Electronic publishing of legislation — towards authenticity
Aki Hietanen ........................................................................................................... 97
Coherence of terminology and search functions
Gianmaria Ajani .................................................................................................... 129
EUR-Lex: from data structures to legal ontologies
Erich Schweighofer ................................................................................................. 137
text mining
Holger Bagola......................................................................................................... 151
PRESS REVIEw / REVUE DE PRESSE
Le quotidien, 30.11.2006 .................................................................................... 172
Luxemburger wort, 24.11.2006 .......................................................................... 173
La Voix, 24.11.2006 ............................................................................................... 173
B.I.t. online 10 (2007) Nr. 1 ................................................................................ 174
Commission en direct, 8.12.2006 ....................................................................... 177
01_2007_5222_txt_ML.indd 10 6-12-2007 15:13:16

En tant que déléguée de la Grèce au groupe de travail “Informatique juridique”
et ayant suivi régulièrement les travaux de ce groupe au sein du Conseil
des ministres, je me réjouis d’avoir la possibilité de fêter les 25 ans du droit
communautaire en ligne et plus de 25 ans d’existence de ce groupe et surtout
de son travail fructueux et ACADEMIC eficace. SESSION
C’était en 1983 que mon pays, la Grèce, se réjouissait du nouveau système
SÉANCE ACADÉMIQUE
Celex, prometteur et qui nous avait ébloui. J’écrivais alors à l’époque, dans une
revue juridique grecque , combien il était nécessaire de disposer d’un système
23 NOVEMBER – 23 NOVEMBRE
donnant accès à toute l’information ju ridique et judicaire, tant pour les services
publics et les universités que pour les particuliers qui s’intéressent au droit
et à son application. Cette nécessité de recherche documentée est devenue
encore plus aiguë en droit communautaire vu le nombre d’actes adoptés, la
diversité des secteurs couverts, la nature spéciale de l’ordre juridique de chaque
état membre, ainsi que la diversité des langues oficielles.
Par sa résolution du 26 novembre 1974 concernant la documentation juridique
automatique, le Conseil des ministres de la Justice avait lancé l’idée de
la création d’un système communautaire de documentation juridique automatique
ouvert à tous les états membres, puisque ceux-ci devaient appliquer le
droit communautaire.
Dans l’article, je me référais aussi au champ couvert par le système, qui
concernait non seulement les textes des traités, mais également le droit dérivé,
les accords conclus par l’Union européenne (à l’époque la Communauté européenne),
la jurisprudence de la Cour de justice ou encore les actes préparatoires.
Ce qui était étonnant, c’était que le système s’étendait aussi aux actes
adoptés par les états membres pour mettre en œuvre le droit communautaire
au sein de leur ordre juridique interne et mentionnait déjà la jurisprudence
nationale. Le système incluait des actes depuis 1952 et la mise à jour se faisait
avec un délai de 3 à 4 semaines à partir de la publication de l’acte. Chaque institution
communautaire était chargée des secteurs le concernant. Le Conseil
fournissait ainsi des données concernant les traités, les relations extérieures, le
droit dérivé, le droit complémentaire. La Commission, le Parlement européen
et le Comité économique et social fournissaient les actes préparatoires. La
HÉMICYCLE KIRCHBERG
Cour fournissait la jurisprudence et le Parlement les questions parlementaires.
La procédure de chargement débutait LUXEMBOURG
par un formulaire en deux parties.
Une partie, assez détaillée, comprenait plus de 30 rubriques de données documentaires,
et une deuxième présentait le texte entier ou un résumé de celui-ci,
selon le schéma du formulaire. On pourrait mentionner certains types de ru-
01_2007_5222_txt_ML.indd 11 6-12-2007 15:13:16

01_2007_5222_txt_ML.indd 12 6-12-2007 15:13:16

MOT DE BIENVENUE
Altesse Royale,
Par Votre présence à notre séance académique de ce jour, une fois de plus,
Vous manifestez Votre engagement et Votre soutien ainsi que ceux du
Grand-Duché de Luxembourg à la construction et au fonctionnement de
l’Union européenne.
Pour ne parler que de cette année, vous avez reçu les présidents des institutions
et des organes européens au palais grand-ducal en début d’année et,
par la suite, lors des manifestations célébrant la fête nationale luxembourgeoise
au mois de juin.
Une telle attention de la part du souverain du pays d’accueil des institutions
est un honneur pour nous, mais surtout un signe tangible et réel de
l’engagement du peuple luxembourgeois et de l’état luxembourgeois en faveur
du mouvement européen, du rapprochement entre ces pays et de l’ancrage
des valeurs de justice, de sécurité et de démocratie dans nos sociétés et
au-delà.
Au nom du comité de direction de l’Ofice des publications et de l’ensemble
du personnel de l’Ofice, je Vous remercie de Votre soutien, qui constitue
également une reconnaissance de la sagesse et de la perspicacité de nos prédécesseurs.
C’est pour leur rendre hommage que nous sommes réunis aujourd’hui.
Altesse Royale,
mesdames et messieurs les membres de la Cour de justice, de la Cour des
comptes et du tribunal de première instance,
monsieur le Doyen du corps diplomatique au Luxembourg, mesdames et
messieurs les Ambassadeurs, mesdames et messieurs les Représentants des
représentants permanents auprès des institutions à Bruxelles,
Chers amis du monde académique et universitaire,
Chers amis des services publics de tous les états membres,
Chers collègues et amis de toutes les institutions européennes,
Chers collègues de l’Ofice,
THOMAS L. CRANFIELD
Directeur de l’Office
des publications
12 | 13
01_2007_5222_txt_ML.indd 13 6-12-2007 15:13:16

Cette énumération des participants à la séance académique de cet aprèsmidi
illustre à quel point le droit européen est devenu une partie centrale de
notre vie de tous les jours. Nous avons l’immense privilège d’être les héritiers de
deux traditions qui sont en train de se fondre dans notre monde changeant.
D’une part, l’Ofice des publications, créé dans le cadre de la fusion des
Commissions de la Communauté du charbon et de l’acier, de l’Euratom et de
la Communauté économique lors des accords de 1967, et qui, depuis 1969,
publie au nom de toutes les institutions européennes le droit européen et ses
documents préparatoires, d’abord sur papier et, par la suite, également sur CD-
ROm et en ligne; et, d’autre part, le développement continu de l’informatique
depuis le début des années 70.
La Commission européenne a eu la grande sagesse de laisser le champ libre
à un groupe d’intellectuels, des juristes éclairés, qui se sont plongés dans ce
nouveau monde technologique et qui, dans les circonstances qui nous seront
racontées tout à l’heure, ont préparé l’Europe aux déis des technologies nouvelles
sur les plans législatif et juridique, des domaines cruciaux pour nos
concitoyens et pour nos sociétés.
Il s’agit en quelque sorte d’un groupe de «prophètes» qui ont vu l’avenir et
qui se sont engagés dans le projet CELEX pendant des années. Les fondations
qu’ils ont mises en place sont toujours là et EUR-Lex de nos jours repose sur
cette base solide, bien pensée et indispensable.
Partant de la vision de quelques-uns, EUR-Lex sert aujourd’hui entre
120 000 et 150 000 citoyens tous les jours. Il s’agit d’un outil iable utilisé dans
tous nos ministères dans tous les pays, dans toutes nos chambres de commerce,
dans les services juridiques des grandes sociétés du secteur privé et
dans les cabinets d’avocats, et ce dans le monde entier.
Avec 1,8 million de documents dans les vingt et une langues oficielles de
l’Union de nos jours — bientôt vingt-trois —, CELEX et maintenant EUR-Lex
est un service public unique au monde et constitue une pierre angulaire de
l’Europe de nos jours.
Je voudrais rendre hommage aux «prophètes» de la première heure, à tous
leurs successeurs et aux entreprises privées qui contribuent à ces travaux tous
les jours.
Cet après-midi, le gouvernement de la finlande, qui assure actuellement la
présidence de l’Union européenne, a accepté de présider notre séance académique,
et je tiens à en remercier m me Rissanen.
m me Durand, qui représente le président Barroso, nous parlera des déis
actuels et ensuite m me Bernet, qui a porté CELEX sur les fonts baptismaux avec
ses collègues de l’époque, nous parlera des débuts de cette aventure.
01_2007_5222_txt_ML.indd 14 6-12-2007 15:13:16

ACADEMIC SESSION /
SÉANCE ACADÉMIQUE
Le professeur herberger décrira les conditions actuelles de l’accès au droit
et esquissera l’une ou l’autre de ses visions pour l’avenir.
m me Rissanen clôturera la séance et nous parlera de l’avenir sur la base de
l’expérience inlandaise.
Altesse Royale,
merci du fond du cœur d’être parmi nous pour ce tour d’horizon.
Chers collègues et amis,
Je vous remercie toutes et tous de votre présence, qui nous encourage à
poursuivre les travaux dans l’intérêt des citoyens et de la construction européenne.
14 | 15
01_2007_5222_txt_ML.indd 15 6-12-2007 15:13:17

01_2007_5222_txt_ML.indd 16 6-12-2007 15:13:17

ALLOCUTION D’OUVERTURE
Altesse Royale,
madame la Secrétaire permanente,
Excellences,
monsieur le Directeur général,
mesdames, messieurs,
CLAIRE-FRANÇOISE DURAND
Directeur général adjoint
à la Commission européenne
(service juridique)
est pour moi un plaisir et un honneur d’introduire ce séminaire pour
C’ fêter les vingt-cinq ans du droit européen en ligne, vingt-cinq ans que le
droit communautaire et ses bases de données sont accessibles au citoyen de
façon aisée et gratuite.
C’est en effet d’abord un plaisir personnel puisque le droit européen en ligne
est né au service juridique de la Commission, de «l’entêtement acharné de
quelques-uns», comme l’écrit m. Cranield dans son livre anniversaire de cet
événement. Et derrière cette expression de m. Cranield, il y a une personne,
l’initiatrice de ce projet, alors qu’elle était au service juridique de la Commission,
m me hélène Bernet. Nous avons la chance de l’avoir parmi nous aujourd’hui
et personne mieux qu’elle ne peut nous raconter les embûches et les joies qui
ont parsemé le chemin de l’élaboration de ce superbe outil, mis à la disposition
du public le 1 er juillet 1981, date dont nous célébrons l’anniversaire. Un outil
perfectionné, rafiné et toujours en mouvement depuis.
C’est aussi un honneur de m’adresser à une telle assemblée d’élite et de
représenter ici le président de la Commission, m. Barroso, que vous aviez invité
à ouvrir ce colloque.
Dans un discours célèbre que le président a prononcé à florence en mars
dernier, il mettait en lumière le rôle essentiel que joue dans l’Union européenne
la règle de droit, expression de la spéciicité du modèle communautaire, d’une
part, élément moteur de la construction européenne et de la cohérence du
projet, d’autre part. Je le cite: «La législation européenne est au cœur de ce qui
fait toute la singularité de l’Union européenne.» Et c’est au nom de cette spéciicité
de l’Union et pour donner à la qualité de la législation et à son accessibilité
une dimension à la hauteur de son importance que la présente Commis-
16 | 17
01_2007_5222_txt_ML.indd 17 6-12-2007 15:13:17

sion a fait de l’initiative «mieux légiférer» ou «Better regulation» une des priorités
centrales de son mandat.
Le droit en ligne fait partie intégrante de l’initiative «mieux légiférer». Il en
est sans doute le précurseur. En effet, cette initiative repose sur une idée de
base simple: la législation est destinée aux citoyens et aux entreprises et elle
doit par conséquent être claire et accessible. C’est son premier objectif. L’initiative
«mieux légiférer» donne corps à un second objectif: la législation doit engendrer
une dynamique et ne pas être un frein.
Si toute législation nécessairement ixe des règles, pour harmoniser les
lois, pour protéger la santé ou l’environnement, une bonne législation doit
comprendre les obligations nécessaires et ne doit pas imposer des charges excessives
ou inutiles. Un lien étroit est établi entre une bonne législation et son
impact inancier et économique sur les citoyens, les entreprises, son impact sur
la compétitivité à l’intérieur de la Communauté et sur la capacité concurrentielle
à l’extérieur de la Communauté. Ce second objectif constitue l’élément
nouveau de l’initiative «mieux légiférer».
Ces deux objectifs sous-tendent les modalités de mise en œuvre de l’initiative
«mieux légiférer». Elle vise l’ensemble du processus d’élaboration de la
loi: la nécessité de légiférer, le contenu de la loi, sa qualité rédactionnelle et son
accessibilité. Elle englobe autant la législation nouvelle que la législation
existante. Je développerai brièvement «mieux légiférer» et la qualité de la loi,
«mieux légiférer» et l’accessibilité au droit.
I. «mIEUX LéGIféRER» Et LA qUALIté DE LA LOI
Dans le cadre de la stratégie «mieux légiférer», la qualité de la loi se mesure
à l’aune de trois paramètres.
1) Le premier est sa valeur ajoutée. Par déinition, toute législation devrait
apporter une valeur ajoutée. Au niveau communautaire, il s’agit
même d’une obligation juridique qui résulte du principe de subsidiarité.
toute législation communautaire doit engendrer une valeur ajoutée par
rapport à des législations que les états membres pourraient prendre
eux-mêmes. Et au-delà même du respect de ce principe, tout processus
législatif doit commencer par une analyse de l’utilité et de la nécessité
de légiférer.
2) Le deuxième paramètre est lié au premier. Pour pleinement apprécier la
valeur ajoutée d’une législation, il convient de procéder à une analyse
approfondie de l’impact que la loi envisagée pourrait avoir. Et c’est à
01_2007_5222_txt_ML.indd 18 6-12-2007 15:13:17

ACADEMIC SESSION /
SÉANCE ACADÉMIQUE
cet égard que le processus d’élaboration des lois, mis en place par la
Commission, connaît les avancées les plus nouvelles.
toute législation importante fait désormais au préalable l’objet d’une
analyse d’impact, de ses incidences économiques, sociales, environnementales,
de ses coûts pour les entreprises, par l’intermédiaire d’études
et d’analyses et de vastes consultations des intéressés. Cela permet
d’avoir une image aussi exacte que possible des besoins, des effets bénéiques
ainsi que des effets négatifs de tels ou tels types de législation,
en d’autres termes d’identifier la bonne réponse législative à des besoins
donnés.
3) Le troisième paramètre s’applique au stade de l’élaboration de son
contenu précis: pour assurer une législation utile et accessible, il
convient d’élaborer des lois simples ne contenant que les obligations
nécessaires. Le respect strict de cette exigence pour les nouvelles législations
est mis en œuvre notamment en liaison avec l’analyse d’impact.
Au-delà de cela, la Commission a lancé un important processus de révision
de tout l’acquis communautaire, dans le but de moderniser la législation
existante, la mettre à jour et surtout la simpliier. Cela s’est traduit par la mise
au point d’un programme de simplification de la législation. Le programme
pour les années 2007 à 2009 comprend une centaine d’initiatives concernant
environ 220 actes législatifs. Il a été adopté par la Commission et présenté au
Parlement le 14 novembre dernier dans le cadre d’un paquet de documents sur
«mieux légiférer», dont je recommande la lecture — en ligne, bien sûr.
L’un de ces documents portait plus spéciiquement sur la réduction des
coûts administratifs engendrés par la législation, une ambition dont on parle
depuis quelques temps. Ce document commence à lui donner corps. L’objectif
est de parvenir à une réduction de 25 % des coûts administratifs engendrés par
la législation communautaire ou celle des états membres. De tels coûts peuvent
résulter par exemple d’obligations d’information, de rapports, de certiications,
de formulaires administratifs, exigés actuellement par les législations et
qui pourraient être supprimés. Les premiers jalons de cette tâche ambitieuse
sont donc lancés.
La réduction des coûts administratifs va constituer un des éléments centraux
de la simpliication de l’acquis. mais la simpliication de l’acquis vise
aussi à améliorer l’accessibilité du droit communautaire. J’aborde ici mon second
point: l’accessibilité du droit.
18 | 19
01_2007_5222_txt_ML.indd 19 6-12-2007 15:13:17

II. «mIEUX LéGIféRER» Et L’ACCESSIBILIté DU DROIt
La qualité de la législation et son accessibilité sont évidemment intimement
liés. Il est sans doute même artiiciel de les distinguer. Je traiterai toutefois
brièvement sous le vocable «accessibilité» d’abord de la qualité rédactionnelle,
ensuite de la codiication, enin de l’accès au droit.
1) La qualité rédactionnelle
Nos institutions, Parlement, Conseil et Commission, se sont dotées de
règles de technique législative reprises dans un accord interinstitutionnel sur la
qualité rédactionnelle du 22 décembre 1998.
Cela met en lumière l’importance attachée au niveau des institutions communautaires
à la qualité rédactionnelle des actes, dont l’objectif est de garantir
la sécurité juridique mais surtout de permettre une lecture aisée pour les citoyens,
les entreprises, les administrations qui doivent les appliquer ou les
faire respecter. Je dois ici rendre hommage aux juristes-réviseurs de nos trois
institutions chargés de mettre en œuvre ces règles et d’assurer la qualité rédactionnelle
de la législation. L’incidence de leur travail sur la structure des actes,
sur la rédaction claire et précise de chaque disposition est incomparable. hommage
doit aussi être porté aux services de traduction de nos institutions qui par
leur professionnalisme assurent la disponibilité des textes de loi dans 23 langues
bientôt. Et je suis sûre que ceux dans cette assemblée qui ont la charge de
publier la loi ou le devoir ou le plaisir de la lire peuvent apprécier l’excellence
de leur travail.
La qualité rédactionnelle de chaque acte et de sa traduction constituent à
l’évidence le dernier paramètre de la qualité de la législation et le premier
maillon pour garantir l’accessibilité du droit.
2) La codiication
Le deuxième maillon concerne la simplicité de l’environnement réglementaire
et c’est à cela que contribue la codiication.
La codiication des actes vise à intégrer dans l’acte de base les amendements
successifs qui lui ont été apportés. Les actes antérieurs sont alors abrogés
et un nouvel acte codiié est adopté et publié. L’objectif est clairement de
rendre plus accessible le droit communautaire, qui fait l’objet de modiications
successives, ce qui rend la législation éparse.
01_2007_5222_txt_ML.indd 20 6-12-2007 15:13:17

ACADEMIC SESSION /
SÉANCE ACADÉMIQUE
La codiication est une opération à droit constant sans modiication de la
substance. Elle se distingue de la refonte qui intègre dans un même acte des
parties codiiées et des parties nouvelles. Cette technique est de plus en plus
souvent utilisée lorsque de nouvelles modiications doivent être apportées à un
acte.
La Commission a lancé un programme de codiication qui porte actuellement
sur cinq cents actes. La codiication d’une centaine d’actes est déjà achevée
et l’objectif est de inaliser ce programme d’ici in 2008. C’est le service
juridique de la Commission qui en est chargé.
Le processus de codiication commence avec la version consolidée de
l’acte fourni par l’Ofice des publications.
L’acte codiié donne donc valeur juridique à la consolidation et simpliie
l’environnement réglementaire. La consolidation quant à elle garde l’avantage
de fournir au lecteur en ligne un texte intégrant quasi instantanément les nouveaux
amendements adoptés.
3) L’accès au droit
Le dernier maillon de l’accessibilité du droit, et certainement le plus fondamental,
est la mise à disposition du public de la législation et jurisprudence
communautaire, dans la forme la plus aisée. CELEX et maintenant EUR-Lex
permettent aux citoyens d’accéder au droit de façon gratuite et facile, aux entreprises
d’aisément repérer les obligations qui les concernent, aux spécialistes
d’opérer des recherches documentaires poussées. EUR-Lex est la synthèse de
trois éléments, tous complexes: le droit, la technologie et le multilinguisme, et,
dans ce dernier élément, EUR-Lex est un instrument unique: il offre en ligne le
droit, en 23 langues bientôt, à 486 millions de citoyens chacun dans sa langue.
Comme disait le commissaire figel en janvier dernier: «L’OPOCE est le lien
ultime dans la chaîne du processus législatif.» EUR-Lex est la phase ultime de
l’initiative «mieux légiférer». Il est aussi le lien entre la loi, cette spéciicité de
l’Union européenne, et le citoyen.
Je vais ici laisser la parole aux personnes plus qualiiées que moi sur ce
sujet, madame Bernet, monsieur le professeur, madame la Secrétaire permanente.
Je félicite vivement l’Ofice des publications d’avoir organisé cet événement
et vous souhaite un bon séminaire.
20 | 21
01_2007_5222_txt_ML.indd 21 6-12-2007 15:13:17

01_2007_5222_txt_ML.indd 22 6-12-2007 15:13:17

LES RACINES: HISTOIRE DE CELEX,
DE 1963 À 1986
Altesse Royale,
madame la ministre,
monsieur le Directeur général,
mesdames, messieurs,
Je me présente à vous comme une «has been», quarante-trois ans après la
première étincelle, jaillie d’un article de fond dans le journal Le Monde; quarante
et un ans après le premier prototype dans une boîte à chaussures, prototype
que j’avais créé avec des iches en carton à perforations marginales (ce
sont des «machines logiques» en miniature). Ces prototypes m’ont servi pour
étudier, non pas le volume bien sûr, mais la structure profonde du droit communautaire
sous le triple aspect relationnel, sémantique et temporel:
• relationnel: l’analyse des relations entre actes;
• sémantique: utilisation des mots issus automatiquement du Journal oficiel
et de descripteurs (termes normalisés attribués par analyse);
• temporel: déinir un système de dates accessibles en machine pour le
calcul automatique des périodes de validité.
C’était en 1965. La même année, j’ai suscité l’automatisation des nombreux
actes agricoles éphémères (les trois quarts du Journal oficiel). Seules les
parties variables devaient dorénavant être traduites et révisées, les autres étant
générées par ordinateur.
Nous arrivons à l’an de grâce 1967. Cette année est historiquement intéressante.
Intéressante pour l’Europe, car c’est l’année de la fusion des exécutifs
communautaires issus des trois traités constitutifs. Intéressante pour le système
CELEX et moi-même, car je présente en 1967 un projet à mon directeur
général, michel Gaudet. Il m’envoie à l’université de michigan, pour un séjour
d’études dans le cadre d’un programme d’échanges. Ce séjour m’a servi à perfectionner
le dossier d’analyse. mes contacts étaient deux professeurs de droit
versés en logique et en informatique:
• John Jackson, grand spécialiste du GAtt (ancêtre de l’OmC), qui avait
écrit lui-même un programme pour automatiser les citations (système
très utilisé dans les pays de Common law),
HÉLÈNE BERNET
Professeur honoraire,
université libre de Bruxelles.
Directeur honoraire
à la Commission européenne
(service juridique)
22 | 23
01_2007_5222_txt_ML.indd 23 6-12-2007 15:13:18

• et L. E. Allen, spécialiste de logique déontique, logique utilisable aussi
pour un système relationnel, qui avait déjà écrit des programmes sur un
IBm préhistorique.
À mon retour, toujours en 1967, je fus affectée à temps plein à la création
du système, avec l’aide d’une «demi-secrétaire».
Deux ans plus tard, en 1969, grâce à un contrat, tout le droit en vigueur
était en machine et on a pu recevoir le premier listing accordéon en réponse à
des questions formulées sur cartes perforées. C’était une procédure lente et
lourde, mais on opérait déjà en télégestion; par exemple, on interrogeait du
terminal de Bruxelles l’ordinateur central situé au Luxembourg. On opérait
aussi, c’était nouveau, en multiprogrammation, c’est-à-dire que plusieurs utilisateurs
pouvaient se connecter en parallèle.
En 1972, un nouveau saut quantique... Cette année connut deux événements.
Le prototype du Répertoire, édition zéro, a été imprimé. Il comportait
les actes en vigueur, sélectionnés avec l’indication de tout leur historique. Ce
prototype était déjà édité automatiquement dans toutes les langues communautaires
— il y en avait quatre à l’époque, pour six états membres. L’autre
événement fut la livraison de la première «console» permettant d’interroger en
ligne de Bruxelles, en mode dialogué ou mode «question-réponse», l’ordinateur
central. Cette console annonçait l’interrogation décentralisée dans toutes
les directions générales et la création de la fonction de «correspondant CE-
LEX» — puis plus généralement celle d’«information oficer», médiateurs de
l’information en ligne.
Il y eut ensuite:
• la constitution d’une équipe,
• le démarrage de la coopération des institutions,
• la liaison câblée de l’Assemblée parlementaire française, symbole de
l’ouverture du système au public.
J’évoquerai ces sujets très rapidement, la brochure étant plus explicite. Je
dirai d’abord quelques mots sur l’ambiance générale, les atouts et les dificultés
de l’entreprise.
• Ambiance
Un scepticisme initial régnait tant au sein du service juridique de la Commission
que dans les autres institutions. Cette activité n’était prévue ni dans les statuts
ni dans le budget communautaire: les premiers programmeurs avaient été
recrutés sur des postes de commis et le centre informatique de l’Euratom s’appe-
01_2007_5222_txt_ML.indd 24 6-12-2007 15:13:18

ACADEMIC SESSION /
SÉANCE ACADÉMIQUE
lait encore «atelier de mécanographie», travaillant en monoprogrammation avec
mémoires séquentielles (grandes bandes magnétiques) et cartes perforées.
L’«enquête de besoins» au service juridique donna des résultats totalement
contradictoires. Comment se prononcer sur des choix que l’on a peine à imaginer?
toutefois, les circonstances de l’époque et la coniance du directeur général
permirent une marge d’initiative et la réalisation d’un projet à très long
terme. Si le service juridique avait dû anticiper et justiier toutes les étapes du
projet point par point, je ne sais si le système aurait vu le jour…
• Atouts
J’ai eu le privilège d’être à la fois entrepreneur et fonctionnaire, dans un
climat général qui permettait les initiatives pour un projet à long terme. Je me
rappelle la stupéfaction de la nouvelle fonctionnaire, ingénieur issu du secteur
privé, à qui il incombait de rédiger un plan de développement sur vingt ans.
Dans le privé, vingt mois, c’est déjà très bien.
• Dificulté
ASSISTANCE TECHNIQUE:
Une partie des programmes devaient être écrits sur mesure. mais peu de
programmeurs, même ingénieurs, étaient préparés pour la recherche documentaire,
et encore moins pour la recherche juridique multilingue et multidroits.
COMPLEXITÉ DES TEXTES:
La structure du droit communautaire est complexe. Elle comporte plusieurs
niveaux. Voici quelques exemples:
• La pratique législative, très pudique, des états membres, ne permettait
pas de retrouver la iliation communautaire des mesures nationales
d’exécution.
• Déterminer automatiquement les dates de début et de in de validité de
chaque acte demandait un calcul à partir de plusieurs données.
• Une autre complexité, d’ordre technique, était due à la gestion du Journal
oficiel. L’Ofice des publications faisait appel pour le Journal oficiel
à plusieurs imprimeries utilisant différentes techniques. Il fallait en tenir
compte pour récupérer le texte intégral des actes lisible en machine.
L’alimentation du système en texte intégral, dans toutes les langues
communautaires, n’a donc pu se faire que progressivement.
24 | 25
01_2007_5222_txt_ML.indd 25 6-12-2007 15:13:18

La création d’un réseau de coopération fut une longue affaire. Il s’agissait
d’institutionnaliser l’alimentation du système. Les directions générales de la
Commission se sont progressivement impliquées. Il s’agissait tout d’abord de
vériier la base juridique et la durée de vie des actes. Elles prirent progressivement
en charge toute l’analyse. Puis ce fut le tour des autres institutions, et
CELEX devint oficiellement un «système interinstitutionnel».
Autres ouvertures
COOPÉRATION
Une coopération était organisée au sein du Conseil de l’Europe, qui comportait
alors vingt et un états membres. très tôt, la Commission fut associée
aux travaux du groupe «Informatique juridique» de cette grande Europe. Il
s’agissait par exemple d’élaborer des normes communes de formatage pour
favoriser l’échange de données juridiques enregistrées sur bande magnétique.
Il s’agissait aussi d’harmoniser la terminologie relative à l’informatique juridique.
C’est ainsi que fut créé le terme Rechtsinformatik en allemand, alors que
Juristische Datenverarbeitung paraissait plus orthodoxe à certains puristes. Rappelons
que le terme «ordinateur» était encore de création récente en langue française.
OUVERTURE EXTÉRIEURE DE CELEX VIA EURONET
Une connexion par câble a pu être établie avec l’Assemblée parlementaire
française. Ce fut une liaison prototype en attendant Euronet (précurseur de
l’internet). La vocation de service public de CELEX a ainsi été concrétisée. Le
dernier développement fut d’en faire un service gratuit!
CRÉATION DU GROUPE «INFORMATIQUE JURIDIQUE» AU CONSEIL
L’ouverture au public et la mise en machine des mesures nationales d’exécution
du droit communautaire ont suscité l’intérêt du Conseil de ministres.
Un groupe ad hoc d’experts européens fut créé. Siégeant avec les représentants
de la Commission, il a contribué au cours des années à l’extension de
CELEX — mais la transparence en matière de mesures nationales d’exécution
ne fut pas acquise en un jour.
La transition CIRCE
L’époque héroïque de CELEX a pris in lorsque la direction générale du
personnel et de l’administration a pris en charge la gestion du système, dont le
poids informatique grandissait. Cette époque a connu toutefois des turbulen-
01_2007_5222_txt_ML.indd 26 6-12-2007 15:13:18

ACADEMIC SESSION /
SÉANCE ACADÉMIQUE
ces du fait des transferts et conversions sur d’autres logiciels et matériels.
L’équipe a retenu son soufle.
PAR L’OPOCE: DE CELEX À EUR-LEX
L’institutionnalisation — et l’interinstitutionnalisation — fut achevée lorsque
l’Ofice des publications a pris en charge la gestion de CELEX. L’Ofice a
fusionné CELEX et une interface de télécommunication permettant l’accès direct
aux textes des Journaux oficiels. Cet heureux mariage a permis l’interrogation
publique du droit communautaire via l’internet.
Ce fut la in d’une épopée. En 1972, un délégué allemand du groupe «Informatique
juridique» avait conirmé la faisabilité du Répertoire multilingue des
actes en vigueur. Recruté ensuite par la Commission et intégré dans l’équipe
CELEX, il a accompagné le système au Luxembourg. La boucle est bouclée.
Le nombre de personnes connues qui ont contribué à l’aventure CELEX à
des titres divers au cours de la période sous revue dépasse la quarantaine. J’ai
fourni à l’Ofice une liste — non exhaustive. Les contributions anonymes, occasionnelles
(par sous-traitance) ou indirectes (mesures nationales d’exécution)
ne sont pas dénombrables.
que toutes soient remerciées au nom des citoyens européens, car la transparence
juridique est l’un des piliers de la démocratie.
mission accomplie, j’ai quitté la Commission en 1986 et l’enseignement universitaire
en 1995. Je suis donc une «has been». Une fonctionnaire postactive.
Je suis aussi un jeune espoir. J’étais l’an dernier étudiante à Nanterre. J’ai
malheureusement découvert que l’on ne peut cumuler les réductions «junior»
et les réductions «senior». Donc, cette année, j’ai passé mon diplôme universitaire
en «Santé publique et environnement» à la faculté de médecine de Paris-
Sud. mon mémoire? L’ethno-énergétique. ma joie et ma ierté, c’est, par exemple,
de désodoriser les élevages intensifs d’une manière simple, facile à mettre
en œuvre, 100 % naturelle. J’ai assaini en une seule séance le plus gros lisier
porcin de l’Auxois. C’est aussi d’aider les personnes hypersensibles par une
réinformation cellulaire, l’hygiène de l’habitat et des consignes alimentaires.
En matière de santé publique liée à l’environnement, le Grand-Duché de
Luxembourg est un exemple à suivre. La spécialité médicale y est reconnue, et
j’ai lu qu’il existait un hôpital compétent pour les maladies et inirmités liées à
l’environnement. L’appel solennel lancé par deux cents médecins qui ont signé
l’appel de Paris contribue à l’éveil dans ce domaine.
Je vous remercie pour votre aimable attention.
26 | 27
01_2007_5222_txt_ML.indd 27 6-12-2007 15:13:18

01_2007_5222_txt_ML.indd 28 6-12-2007 15:13:18

DER ZUGANG ZUM RECHT
Der Zugang zum Recht war in der Geschichte des Rechts – soweit wir sie
kennen – von Anfang an ein thema.
Betrachten wir dazu drei Stationen aus der Überlieferung.
Die erste Erinnerung ist (ungefähr) mit dem Jahr 1750 vor Christus verbunden.
Im „Codex hammurabi“ rühmt sich der herrscher im Epilog, den Unterdrückten
Zugang zum Recht gegeben zu haben:
‘the king who rules among the kings of the cities am I. my words are well
considered; there is no wisdom like mine. By the command of Shamash,
the great judge of heaven and earth, let righteousness go forth in the land:
by the order of marduk, my lord, let no destruction befall my monument.
In E-Sagil, which I love, let my name be ever repeated; let the oppressed,
who have a case at law, come and stand before this my image as king of
righteousness; let him read the inscription, and understand my precious
words: the inscription will explain his case to him; he will ind out what is
just, and his heart will be glad, so that he will say:
“hammurabi is a ruler, who is as a father to his subjects, who holds the
words of marduk in reverence, who has achieved conquest for marduk
over the north and south, who rejoices the heart of marduk, his lord, who
has bestowed beneits for ever and ever on his subjects, and has established
order in the land.”’
[the Code of hammurabi, Epilogue, translated by L.W. King (1910), edited by Richard
Hooker, http://www.wsu.edu/~dee/mESO/CODE.htm]
„Zugang“ ist hier zunächst wörtlich im Sinne von „Zu-Gang“ gemeint:
Der Rechtsuchende betritt den tempel, in dem sich die Stele mit hammurabis
Codex beindet. Er darf sich dann – dies setzt der text voraus – dieser
MAXIMILIAN HERBERGER ( 1 )
Professor an der Universität
des Saarlandes, Deutschland
( 1 ) Prof. Dr. maximilian herberger ist Direktor des Instituts für Rechtsinformatik an der
Universität des Saarlandes (http://rechtsinformatik.jura.uni-saarland.de/), Vorsitzender des
Deutschen EDV-Gerichtstages e. V. (http://www.edvgt.de/) und herausgeber der freien
Internet-Zeitschrift für Rechtsinformatik und Informationsrecht JurPC (http://jurpc.de/).
28 | 29
01_2007_5222_txt_ML.indd 29 6-12-2007 15:13:19

Stele so weit nähern, dass er den darauf beindlichen text lesen kann. Es soll
sich aber bei dieser Begegnung mehr ereignen als ein bloßes Zurkenntnisnehmen
der worte des Gesetzes. Bemerkenswerterweise wird betont, dass
der sein Recht Suchende bei der Lektüre den text des Gesetzes verstehen
können soll:
‘let him read the inscription, and understand my precious words …’
So meint „Zu-Gang“ am Ende und im vollen Verständnis des wortes nicht
nur den „äußeren“ Zugang zum text, sondern darüber hinaus den „inneren“
Zugang zum Sinn des textes, eine – wie wir sehen werden – folgenreiche Verknüpfung
auch für unser heutiges Verständnis.
Betrachten wir eine zweite Episode, in der „Zugang zum Recht“ ein thema
ist, diesmal aus dem Alten testament. wir beinden uns mit diesem Bericht in
der mitte des 5. Jahrhunderts vor Christus, in der Zeit der Rückkehr des jüdischen
Volkes aus dem Babylonischen Exil.
Im Buch Nehemia lesen wir:
„Als nun der siebente Monat herangekommen war und die Israeliten in ihren Städten
waren, versammelte sich das ganze Volk wie ein Mann auf dem Platz vor dem
Wassertor, und sie sprachen zu Esra, dem Schriftgelehrten, er solle das Buch des
Gesetzes des Mose holen, das der HERR Israel geboten hat. Und Esra, der Priester,
brachte das Gesetz vor die Gemeinde, Männer und Frauen und alle, die es verstehen
konnten, am ersten Tage des siebenten Monats und las daraus auf dem Platz vor
dem Wassertor vom lichten Morgen an bis zum Mittag vor Männern und Frauen
und wer’s verstehen konnte. Und die Ohren des ganzen Volks waren dem
Gesetzbuch zugekehrt. Und Esra, der Schriftgelehrte, stand auf einer hölzernen
Kanzel, die sie dafür gemacht hatten. ... Und Esra tat das Buch auf vor aller Augen,
denn er überragte alles Volk; und als er’s auftat, stand alles Volk auf. Und Esra lobte
den HERRN, den großen Gott. Und alles Volk antwortete: ‚Amen! Amen!‘, und sie
hoben ihre Hände empor und neigten sich und beteten den HERRN an mit dem
Antlitz zur Erde. Und die Leviten ... unterwiesen das Volk im Gesetz, und das Volk
stand auf seinem Platz. Und sie legten das Buch des Gesetzes Gottes klar und
verständlich aus, sodass man verstand, was gelesen worden war.
Und Nehemia, der Statthalter, und Esra, der Priester und Schriftgelehrte, und die
Leviten, die das Volk unterwiesen, sprachen zu allem Volk: Dieser Tag ist heilig dem
HERRN, eurem Gott; darum seid nicht traurig und weinet nicht! Denn alles Volk
weinte, als sie die Worte des Gesetzes hörten. Darum sprach er zu ihnen: Geht hin
und esst fette Speisen und trinkt süße Getränke und sendet davon auch denen, die
nichts für sich bereitet haben ... Und alles Volk ging hin, um zu essen, zu trinken und
01_2007_5222_txt_ML.indd 30 6-12-2007 15:13:19

ACADEMIC SESSION /
SÉANCE ACADÉMIQUE
davon auszuteilen und ein großes Freudenfest zu machen; denn sie hatten die Worte
verstanden, die man ihnen kundgetan hatte.“
[Buch Nehemia, Kapitel 8, Vers 1-12;
http://www.bibel-online.net/buch/16.nehemia/8.html].
wieder wird – um im Duktus unserer Überlegungen zu bleiben – das Gesetz
dem Volk zugänglich gemacht und nahegebracht, diesmal in der doppelten
form von Rezitation und Kommentierung. Auffällig ist, wie oft in dem
text darauf wert gelegt wird, dass der rezitierte und ausgelegte text auch verstanden
wurde, und dies durch die erste Erschütterung und die tränen hindurch
bis hin zur freude über das Gesetz. Das zeigt, dass der bloße Kontakt
mit dem Gesetz durch Anhören ohne innere Empathie allenfalls als ein Durchgangsstadium
angesehen wird, das durch zusätzliche Bemühungen hin zu
einem wirklichen Verständnis des Gesetzestextes geführt werden muss – auch
dies ein heute noch bedenkenswerter Gedanke.
Die dritte „access story“ spielt im Jahr 304 vor Christus in Rom.
wurde in den bisherigen texten davon erzählt, wie der Zugang zum Recht
gewährt wurde, so erinnert uns diese Geschichte daran, dass der Zugang zum
Recht unter bestimmten Umständen erkämpft werden muss. Diesbezüglich
wird in Digesten 1.2.2.7 folgendes überliefert:
„Pomponius l.S. enchir.
Postea cum Appius Claudius proposuisset et ad formam redegisset has actiones,
Gnaeus Flavius scriba eius libertini filius subreptum librum populo tradidit,
et adeo gratum fuit id munus populo, ut tribunus plebis fieret et senator et
aedilis curulis. Hic liber, qui actiones continet, appellatur ius civile flavianum, sicut
ille ius civile papirianum: nam nec Gnaeus Flavius de suo quicquam adiecit libro.“
„Später hat, nachdem Appius Claudius diese Formeln vorgelegt und in eine neue
Form gebracht hatte, dessen Schreiber Gnaeus Flavius, der Sohn eines Freigelassenen,
das Buch gestohlen und dem Volk übergeben und das Volk war dafür so dankbar,
dass er Volkstribun und Senator und kurulischer Edil wurde. Dieses Buch, das die
Klageformeln enthält, wird das flavianische Zivilrecht genannt, so wie jenes andere
Buch, das papirianische Zivilrecht. Denn auch Gnaeus Flavius hat dem Buch nichts
Eigenes hinzugefügt.“
[Übersetzung von thomas Rüfner, http://www.uni-trier.de/~ruefner/Lehre/RR-
Geschichte/RR-AB2.pdf ; dort auch die Übersetzung des Berichts von Livius (Ab
urbe condita 9, 46, 1-9) zum gleichen thema.]
was war geschehen? Der Schreiber Gnaeus flavius, Sohn eines freigelassenen,
der im Dienst des Patriziers Appius Claudius Caecus stand, hatte eine
30 | 31
01_2007_5222_txt_ML.indd 31 6-12-2007 15:13:19

nicht autorisierte Abschrift der ofiziellen Klagformeln und des Gerichtskalenders
angefertigt und veröffentlicht. Das führte dazu, dass das wissen um das
Recht und die tage der Gerichtsverhandlungen nicht mehr exklusiv den Priesterkollegien
zur Verfügung stand, sondern für das Volk zugänglich geworden
war. Das Volk dankte ihm diese tat durch Ämterverleihungen; es muss also
wohl den Zugang zum Recht als wohltat empfunden haben. Damit verbunden
war sicher der Glaube, das so zugänglich gewordene Recht verstehen zu
können und verstanden zu haben, denn sonst hätte man den Zugriff darauf
nicht dankbar als wohltat empfunden. wiederum begegnen wir somit dem
Zusammenklang von Zugänglichwerden des Rechts und darauf gerichteten
Verständlichkeitserwartungen – dies alles begleitet von guten Gefühlen.
Diese drei Rückblicke zeigen, dass der Zugang zum Recht im Spannungsfeld
von Gewährung und Kampf immer schon ein thema war – so wie heute
(und hier und heute) auch für uns.
wagen wir also den Sprung in die Gegenwart und gratulieren zunächst all
denen, die vor 25 Jahren begonnen haben, das Europäische Recht online den
Bürgerinnen und Bürgern Europas zugänglich zu machen, und all denen, die
diese Initiative bis heute lebendig gehalten haben. Bei meinen Streifzügen im
Internet zur Vorbereitung dieses Vortrags iel mir auf, dass eine australische
Universität aus festlichem Anlass einmal CELEX als Akronym für „CELebrating
EXcellence“ verwandt hat. wir dürfen die marke CELEX heute auch so
deuten, denn es gilt, einer exzellenten Leistung Reverenz zu erweisen.
Aber, eine weiter gehende frage muss auch an einem festlichen tag wie
diesem erlaubt sein:
Ist es damit getan, die texte des Rechts in leicht erreichbarer form der
Öffentlichkeit (früher hätte man gesagt „dem Volk“) zur Verfügung zu stellen?
Oder gibt es eine darüber hinausreichende Zukunftsaufgabe, will man das Programm
„Zugang zum Recht“ in einem vertieften Sinn weiterführen?
Gewiss: Der erste Schritt, Zugang zu den texten zu verschaffen – und dies
kostenfrei – war und ist notwendig.
Aber – diese these soll jetzt verfochten werden – es ist dies nur eine notwendige,
keine hinreichende Bedingung für den Zugang zum Recht im Vollsinn
des wortes. Denn können wir wirklich sagen, das Recht sei „zugänglich“
geworden, wenn Bürgerinnen und Bürger mit den für sie erreichbaren texten
konfrontiert diese für unverständlich halten? was wäre dadurch anders gewonnen
als der Zugang zu einem Arkanum, das sich nach wie vor als unzugänglich-hermetisch
erweist? Und was wären die Konsequenzen davon?
01_2007_5222_txt_ML.indd 32 6-12-2007 15:13:19

ACADEMIC SESSION /
SÉANCE ACADÉMIQUE
Übrigens kommt hilfe für die these, Zugang zum Recht im Vollsinn setze
ein prinzipiell verständliches Recht voraus, von unerwarteter Seite. Sicher haben
Sie schon von den „web Accessibility Guidelines“ des w3-Konsortiums
gehört, die Regeln für die Zugänglichkeit von websites aufstellen.
[Ich zitiere auf Englisch, weil nur der englische text verbindlich ist:
‘the English version of this speciication is the only normative version.’
[http://www.w3.org/tR/wCAG10/]
Dort lesen wir in Guideline 14:
‘Ensure that documents are clear and simple so they may be more
easily understood.
Consistent page layout, recognisable graphics, and easy to understand
language beneit all users. …
Using clear and simple language promotes effective communication.
Access to written information can be dificult for people who have
cognitive or learning disabilities. Using clear and simple language also
beneits people whose irst language differs from your own, including
those people who communicate primarily in sign language.’
[http://www.w3.org/tR/wCAG10/#gl-facilitate-comprehension]
Anschließend werden „checkpoints“ formuliert. Checkpoint 14.1 lautet:
‘14.1. Use the clearest and simplest language appropriate for a site’s
content. [Priority 1]’
[http://www.w3.org/tR/wCAG10/]
Dieser Checkpoint ist mit Priorität 1 versehen. Und das bedeutet:
‘[Priority 1]
A web content developer must satisfy this checkpoint. Otherwise, one or
more groups will ind it impossible to access information in the document.
Satisfying this checkpoint is a basic requirement for some groups to be
able to use web documents.’
[http://www.w3.org/tR/wAI-wEBCONtENt-tEChS/#tech-simple-andstraightforward]
was wir hier lesen, ist z. B. in Deutschland für die Internet-Auftritte von
Behörden der Bundesverwaltung geltendes Recht. Denn der Verordnungsgeber
hat sich entschlossen, in der BItV (der „Verordnung zur Schaffung barrierefreier
Informationstechnik nach dem Behindertengleichstellungsgesetz“) im dortigen Anhang
folgendes zu verordnen:
32 | 33
01_2007_5222_txt_ML.indd 33 6-12-2007 15:13:19

„Die Anforderungen und Bedingungen dieser Anlage basieren grundsätzlich auf den
Zugänglichkeitsrichtlinien für Web-Inhalte 1.0 (Web Content Accessibility Guidelines
1.0) des World Wide Web Consortiums vom 5. Mai 1999.“
[http://www.gesetze-im-internet.de/bitv/BJNR265400002.html]
Und da das deutsche Bundesministerium der Justiz das Bundesrecht im
Internet veröffentlicht, müsste es als „web content developer“ für die Inhalte
dieser website darauf achten, dass die Sprache der dort publizierten Gesetze
„clear and simple“ ist ...
Aber sprechen wir überhaupt von einem wirklichen Problem? Ist mit Blick
auf das aktuelle EU-Recht das Problem der mangelnden Verständlichkeit wirklich
so ernst, dass man von einem schwerwiegenden Zugangshindernis sprechen
darf?
Ich möchte hier – neben Ihrer eigenen Erfahrung mit der materie – nur eine
Kronzeugin aufrufen, die Schweizer Bundeskanzlerin Annemarie huber-hotz.
Sie hat im Jahr 2000 zu diesem thema folgendes gesagt:
„Die größte Herausforderung für die Verständlichkeit unserer Gesetze stellt aber zurzeit
zweifellos das EU-Recht dar, das wir seit längerem ‚autonom nachvollziehen‘ und mit
den bilateralen Abkommen zum Teil nun auch direkt anwenden. Verstehen Sie mich
richtig: Ich bin eine überzeugte Europäerin, aber ich meine, wir sollten die manchmal
geradezu groteske Unüberschaubarkeit, Umständlichkeit und Unverständlichkeit der
EU-Rechtserlasse nicht einfach als ein Naturgesetz hinnehmen. Als kleines Muster
gebe ich Ihnen nur einen Titel einer EU-Verordnung, und es ist mir durchaus bewusst,
dass ich Ihnen jetzt etwas zumute.
Der Titel lautet:
‚Verordnung (EG) Nr. 2592/1999 der Kommission vom 8. Dezember 1999 zur
Änderung der Verordnung (EG) Nr. 1826/1999 zur Änderung der Verordnung (EG)
Nr. 929/1999 zur Einführung vorläufiger Antidumping- und Ausgleichszölle auf die
Einfuhren von gezüchtetem Atlantischen Lachs mit Ursprung in Norwegen im Fall
bestimmter Ausführer, zur Einführung vorläufiger Antidumping- und Ausgleichszölle
auf die Einfuhren von solchem Lachs im Fall bestimmter Ausführer, zur Änderung des
Beschlusses 97/634/EG zur Annahme von Verpflichtungsangeboten im Zusammenhang
mit dem Antidumping- und dem Antisubventionsverfahren betreffend die Einfuhren von
solchem Lachs und zur Änderung der Verordnung (EG) Nr. 772/1999 des Rates zur
Einführung endgültiger Antidumping- und Ausgleichszölle auf die Einfuhren von
solchem Lachs‘ [ABl. L 315 vom 9.12.1999, S. 17]“
[Annemarie huber-hotz, Recht haben – gerecht sein, http://web.archive.org/
web/20030701101153/http://www.admin.ch/ch/d/bk/hu20001106.html]
01_2007_5222_txt_ML.indd 34 6-12-2007 15:13:19

ACADEMIC SESSION /
SÉANCE ACADÉMIQUE
Übrigens ergibt sich (nebenbei bemerkt) aus der tatsache einer möglicherweise
unverständlichen EU-Verordnung eine interessante europarechtliche
frage:
Darf man eine (an Verständlichkeitsmängeln leidende) Gemeinschaftsverordnung
aus Gründen der Verständlichkeit in einzelstaatliches Recht umsetzen,
obwohl sie dort schon von vornherein (aber möglicherweise in unverständlicher
weise) ohne Umsetzung gilt?
Die Kommission hat dies in einer an Dänemark gerichteten Empfehlung
aus dem Jahr 1973 für möglich erachtet:
„Die Vorschriften dieser Verordnung gelten aufgrund von Artikel 189 des Vertrages
unmittelbar in den Mitgliedstaaten und sind in Dänemark am 1. Januar 1973 in Kraft
getreten. Folglich sind sie nicht in das einzelstaatliche Recht zu übernehmen. Eine derartige
Übernahme könnte im Übrigen zu Missverständnissen hinsichtlich des Inkrafttretens und
der unmittelbaren Anwendbarkeit derartiger Gemeinschaftsverordnungen führen. Eine
auch nur teilweise Übernahme von Vorschriften einer Gemeinschaftsverordnung
in einzelstaatliches Recht wäre höchstens zulässig, wenn sie zur Verständlichkeit
erforderlich wäre.“
[73/157/EwG: Empfehlung der Kommission vom 27. April 1973 an die Regierung
Dänemarks zu dem Entwurf eines Erlasses zur Durchführung der Verordnung Nr.
117/66/EwG sowie der Verordnung (EwG) Nr. 1016/68, Amtsblatt L 182 vom
5.7.1973, S. 27, http://europa.eu.int/smartapi/cgi/sga_doc?smartapi!celexapi!prod!
CELEXnumdoc&lg=DE&numdoc=31973h0157&model=guichett ]
Da bleibt also in bemerkenswerter weise Raum für dem europäischen
Recht nachgehende nationale Verständlichkeitsanstrengungen.
Ich bin kein Utopist, der meint, man könne unter heutigen Bedingungen das
Recht insgesamt für die Bürgerinnen und Bürger „in den Straßen“ verständlich
präsentieren. Aber sollten wir nicht die hoffnung aufrechterhalten, dass dies für
einen Kernbestand und die Grundgedanken möglich sein müsste? Denn wenn
es uns nicht gelingt, diesen Eindruck glaubhaft zu vermitteln, droht auch aus
diesem Grunde ein mentaler Exodus der Bürgerinnen und Bürger aus dem gemeinsamen
Rechtsraum, den die Europäische Union bauen will.
Können wir nach alledem nun sagen „Problem erkannt – Gefahr gebannt?“
Das wird angesichts des Ernstes der Lage wohl nicht ausreichend sein. Aber
was kann man darüber hinaus tun?
Zum einen sollte man mit Beharrlichkeit an all das erinnern, was in der
Europäischen Union diesbezüglich an guten Vorsätzen und guten Beschlüssen
bereits existiert. Zu nennen ist in dieser hinsicht vor allem die „Interinstitutionelle
Vereinbarung vom 22. Dezember 1998 – Gemeinsame Leitlinien für die
redaktionelle qualität der gemeinschaftlichen Rechtsvorschriften“, die das Ziel
34 | 35
01_2007_5222_txt_ML.indd 35 6-12-2007 15:13:19

hat, „Zugänglichkeit und Verständlichkeit der gemeinschaftlichen Rechtsvorschriften“
ständig zu verbessern. Sie bringt „Zugang“ und „Verständlichkeit“
in den auch hier befürworteten notwendigen Zusammenhang und nennt als
obersten Grundsatz:
„1. Die gemeinschaftlichen Rechtsakte werden klar, einfach und genau
abgefasst.“
(Amtsblatt C 73 vom 17.3.1999, S. 1-4, http://europa.eu.int/eur-lex/lex/
LexUriServ/LexUriServ.do?uri=CELEX:31999y0317(01):DE:htmL)
Und zum anderen sollte man mit ein wenig Phantasie über neue formen
der Vermittlung von Recht nachdenken.
warum z. B. nicht dem Vorschlag von martin Cutts folgen? Er befürwortet
eine den Richtlinien vorangestellte Zusammenfassung für den Bürger:
„Wenige Menschen machen sich die Mühe, eine Richtlinie zu lesen. Einer der Gründe
ist ihre Fremdartigkeit; sie gleichem keinem anderen zuvor gelesenen Text. Richtlinien
haben nur wenige Merkmale, die normalen Bürgern entgegenkommen; diese finden
sich jäh auf nicht vertrautem Gelände. Ein Weg zur Lösung dieses Problems besteht
darin, an den Beginn jeder Richtlinie eine kurze Zusammenfassung der wesentlichen
Punkte zu stellen – eine Zusammenfassung für den Bürger.“
[http://209.85.135.104/search?q=cache:kswvCrf-O-UJ:europa.eu.int/comm/
translation/en/ftfog/eurolaw/eurolaw_de.htm+site:europa.eu.int+Verst%C3%A4
ndlichkeit&hl=en&ct=clnk&cd=41 ]
Oder, wenn man noch Innovativeres erwägen will: warum nicht visuelle
formen der Vermittlung von Recht erproben? wenn es wahr ist, dass ein Bild
mehr als tausend worte sagen kann, resultieren vielleicht einige unserer
Schwierigkeiten aus einem bloß textbezogenen Denken.
Schließen möchte ich mit einer Geschichte und einer frage.
Die Geschichte kann ich nicht besser erzählen, als dies Berkemann 1999
beim 8. Deutschen EDV-Gerichtstag in seinem Vortrag „freies Recht für freie
Bürger“ getan hat. Sie spielt im 18. Jahrhundert und handelt von Johann Georg
Krünitz. Nach Krünitz ist
„es nötig, ‚dass Landesgesetze dem gemeinen Manne billig durch Kalender
beygebracht werden sollten‘. Ein solcher jährlich erscheinender Polizeikalender hätte
dann die Aufgabe, ‚den Leuten nach und nach den Inhalt der alten und in Observanz
stehenden sowohl als neuen Polizey- und Landes-Gesetze und Verordnungen
auszugsweise, und durch Erinnerungen und Anmerkungen, bekannt zu machen,
und ihnen dadurch mit der Zeit ganze Sammlungen derselben in die Hände zu
bringen‘.
01_2007_5222_txt_ML.indd 36 6-12-2007 15:13:20

ACADEMIC SESSION /
SÉANCE ACADÉMIQUE
Krünitz meinte, dieser Polizeikalender sei das wirksamste Mittel, eine umfassende
Gesetzeskenntnis zu erreichen. Die Erwachsenen, die Kinder und das Gesinde
würden die Texte ‚oft, und wenn es auch nur zur Verkürzung der Zeit geschehen
sollte‘, lesen und diesem dadurch ‚zuletzt vollkommen bekannt werden‘. – Man sieht
es gerade vor sich: Der Familienvater – die jüngste Sendung des Polizeikalenders in
den Händen – liest der Familie in abendlicher Runde die neuen Rechtsnormen vor,
vergewissert sich, ob die Kenntnisse aus früheren Tagen noch vorhanden sind.“
[Jörg Berkemann, freies Recht für freie Bürger! JurPC web-Dok. 188/1999, Abs. 1-
79, http://www.jurpc.de/aufsatz/19990188.htm ]
So wird es nicht wieder werden (wenn es denn jemals so war).
Aber wie stellen wir es uns heute vor?
Das ist immer noch die zentrale frage für unser Verständnis von „Zugang
zum Recht“.
[TRANSLATION INTO ENGLISH]
Access to law
Access to law has been an issue throughout
legal history as we know it. Let us consider
three historical milestones.
the irst dates from around 1 750 BC. In the
epilogue to the ‘Code of hammurabi’, the
ruler boasts about having given the oppressed
access to law:
‘the king who rules among the kings of
the cities am I. my words are well considered;
there is no wisdom like mine. By the
command of Shamash, the great judge of
heaven and earth, let righteousness go
forth in the land: by the order of marduk,
my lord, let no destruction befall my
monument. In E-Sagil, which I love, let
my name be ever repeated; let the oppressed,
who have a case at law, come and
stand before this my image as king of
righteousness; let him read the inscription,
and understand my precious words: the
inscription will explain his case to him; he
will ind out what is just, and his heart will
be glad, so that he will say:
“hammurabi is a ruler, who is as a father
to his subjects, who holds the words of
marduk in reverence, who has achieved
conquest for marduk over the north and
south, who rejoices the heart of marduk,
his lord, who has bestowed beneits for
ever and ever on his subjects, and has established
order in the land.”’
[the Code of hammurabi, Epilogue, translated by
L. W. King (1910), edited by Richard Hooker (http://
www.wsu.edu/~dee/mESO/CODE.htm)]
In German, the word for ‘access’ is ‘Zugang’,
which means literally ‘going in’. And we see
this here — the person seeking justice ‘goes
in’ to the temple where the stele on which the
Code of hammurabi is written can be found.
the text goes on to say that he may approach
the stele so that he can read the inscription.
But this is to be more than just reading the
wording. Interestingly, it is emphasised that
the person seeking justice should be able to
understand the text of the law:
‘let him read the inscription, and understand
my precious words …’
So ‘access’ here means not only ‘supericial’
access to the text as a piece of writing, but
also ‘internalised’ access to the meaning of
the text. this is an important distinction for
our present-day understanding, as we will
see.
Let us now turn to a second episode where
‘access to law’ is an issue, this time from the
36 | 37
01_2007_5222_txt_ML.indd 37 6-12-2007 15:13:20

Old testament. this passage is set in the
middle of the ifth century BC, at the time of
the return of the Jews from their exile in Babylon.
In the Book of Nehemiah we read the
following:
‘1. And all the people gathered themselves
together as one man into the street that
was before the water gate; and they spake
unto Ezra the scribe to bring the book of
the law of moses, which the Lord had
commanded to Israel. 2. And Ezra the
priest brought the law before the congregation
both of men and women, and all
that could hear with understanding,
upon the irst day of the seventh month.
3. And he read therein before the street
that was before the water gate from the
morning until midday, before the men
and the women, and those that could
understand; and the ears of all the people
were attentive unto the book of the law.
4. And Ezra the scribe stood upon a pulpit
of wood, which they had made for the
purpose …
5. And Ezra opened the book in the sight
of all the people; (for he was above all the
people;) and when he opened it, all the
people stood up. 6. And Ezra blessed the
Lord, the great God. And all the people
answered, Amen, Amen, with lifting up
their hands: and they bowed their heads,
and worshipped the Lord with their faces
to the ground … the Levites, caused the
people to understand the law: and the
people stood in their place. 8. So they read
in the book in the law of God distinctly,
and gave the sense, and caused them to
understand the reading.
9. And Nehemiah, which is the tirshatha,
and Ezra the priest the scribe, and the Levites
that taught the people, said unto all
the people, this day is holy unto the Lord
your God; mourn not, nor weep. for all
the people wept, when they heard the
words of the law. 10. then he said unto
them, Go your way, eat the fat, and drink
the sweet, and send portions unto them
for whom nothing is prepared: for this
day is holy unto our Lord: neither be ye
sorry; for the joy of the Lord is your
strength … 12. And all the people went
their way to eat, and to drink, and to send
portions, and to make great mirth, because
they had understood the words
that were declared unto them.’
[Nehemiah, Chapter 8, Verses 1–12 (King James
Version)]
Once again, the law was made accessible to
the people and presented to them in dual
form — irst recitation and then commentary.
It is striking how often the passage emphasises
that the recited text was also explained
and understood, and how the people’s initial
reaction of upset and tears gave way to joy.
this shows that mere contact with the law
by hearing it without really internalising it
can be considered a transitional stage, but
that additional explanation is required to lead
to real understanding of the text. this is a
point which is still relevant today.
the third ‘access story’ takes place in Rome
in 304 BC. whilst the irst two texts talk
about access to the law being granted, the
third reminds us that, under certain circumstances,
access to the law is a struggle. Digesta
1.2.2.6–7 tells us the following:
‘Pomponius, Enchiridion
(6) Deinde ex his legibus eodem tempore
fere actiones compositae sunt, quibus inter
se homines disceptarent: quas actiones
ne populus prout vellet institueret certas
solemnesque esse voluerunt: et appellatur
haec pars iuris legis actiones, id est legitimae
actiones. Et ita eodem paene tempore
tria haec iura nata sunt: lege duodecim
tabularum ex his luere coepit ius civile, ex
isdem legis actiones compositae sunt. Omnium
tamen harum et interpretandi scientia
et actiones apud collegium pontiicum
erant, ex quibus constituebatur, quis quoquo
anno praeesset privatis. Et fere populus
annis prope centum hac consuetudine
usus est.
(7) Postea cum Appius Claudius proposuisset
et ad formam redegisset has actiones,
Gnaeus Flavius scriba eius libertini filius
subreptum librum populo tradidit, et
adeo gratum fuit id munus populo, ut
tribunus plebis fieret et senator et aedilis
curulis. hic liber, qui actiones continet,
01_2007_5222_txt_ML.indd 38 6-12-2007 15:13:20

ACADEMIC SESSION /
SÉANCE ACADÉMIQUE
appellatur ius civile lavianum, sicut ille
ius civile papirianum: nam nec Gnaeus
flavius de suo quicquam adiecit libro.’
what had happened? the writer Gnaeus flavius,
son of an emancipated slave, in the
service of the patrician Appius Claudius
Caecus, had drawn up and published an unauthorised
copy of the oficial formulae for
submitting a case and the judicial calendar.
this meant that knowledge of the law and
the dates of court proceedings was no longer
the preserve of the priestly class but had
been made accessible to the people. the
people thanked him for this act by bestowing
ofices on him; they must therefore have
considered providing access to the law to be
a good deed. this implies that the law thus
made accessible was comprehensible and
understood, as otherwise there would have
been nothing to be thankful for. So, once
again, we encounter the link between making
the law accessible and the expectation
that it will be understandable — accompanied
by feelings of satisfaction.
these three historical documents show that
people have always been concerned by access
to the law and the tension between its
being granted and fought for — as is still the
case for us here and now.
Let us then move forward to the present
and congratulate all those who, 25 years
ago, started making European law available
online to Europe’s citizens, and all those
who have kept this initiative going until today.
I was suring the Internet in preparation
for this talk and found that an Australian
university once used the acronym
CELEX — obviously without alluding to
‘our’ CELEX — to mean ‘CELebrating EXcellence’.
It should mean the same to us, as
CELEX is an excellent achievement to be
honoured.
however, even on such an auspicious day,
more far-reaching questions should be allowed,
such as this: Is it enough to make the
texts of the law available to the public (previously,
we would have said ‘the people’) in
easily accessible form? Or should our task
for the future be to expand our interpretation
of ‘providing access to the law’?
It goes without saying that the irst step —
providing access to the texts, free of
charge — was, and still is, necessary.
But — and this is my hypothesis — this step
is just one necessary condition for making
the law accessible in the full meaning of the
word, but it is by no means suficient. Can
we really say that the law is ‘accessible’ if
citizens, when confronted by a text they
have accessed, ind it incomprehensible?
what would be gained if citizens were to
gain access to one ‘secret’ only to be faced
by another equally impenetrable one? And
what would be the consequences?
however, support for my hypothesis that
access to the law implies that this law is
comprehensible can be found from an unexpected
side. I am sure you have heard of the
web Accessibility Guidelines of the world
wide web Consortium, which establish
rules for the accessibility of websites.
‘the English version of this speciication is
the only normative version.’
[http://www.w3.org/tR/wCAG10/]
Guideline 14:
‘Ensure that documents are clear and simple
so they may be more easily understood.
Consistent page layout, recognisable
graphics, and easy to understand language
beneit all users ...
Using clear and simple language promotes
effective communication. Access to written
information can be dificult for people
who have cognitive or learning disabilities.
Using clear and simple language also
beneits people whose irst language differs
from your own, including those
people who communicate primarily in
sign language.’
[http://www.w3.org/tR/wCAG10/#glfacilitate-comprehension]
‘Checkpoints’ are then set out. Checkpoint
14.1 states:
‘14.1. Use the clearest and simplest language
appropriate for a site’s content. (Priority
1)’
[http://www.w3.org/tR/wCAG10/]
38 | 39
01_2007_5222_txt_ML.indd 39 6-12-2007 15:13:20

this checkpoint is given a Priority 1, which
means that:
‘(Priority 1)
A web content developer must satisfy this
checkpoint. Otherwise, one or more
groups will ind it impossible to access information
in the document. Satisfying this
checkpoint is a basic requirement for
some groups to be able to use web documents.’
[http://www.w3.org/tR/wAI-wEBCONtENttEChS/#tech-simple-and-straightforward]
this is the standard in Germany, for example,
for the websites of federal authorities.
the regulator laid down the following
in the annex to the BITV (Verordnung zur
Schaffung barrierefreier Informationstechnik nach
dem Behindertengleichstellungsgesetz — Order
on the creation of accessible It in accordance
with the Equal treatment (Disability)
Act):
‘the requirements and conditions in this
Annex are based, in principle, on the web
Content Accessibility Guidelines 1.0 of
the world wide web Consortium of 5
may 1999.’
[http://www.gesetze-im-internet.de/bitv/
BJNR265400002.html]
As the federal ministry of Justice publishes
federal law on the Internet, it is therefore responsible,
as the web content developer, for
ensuring that the language used in the laws
published there is clear and simple.
But is this a real problem? when it comes to
today’s EU law, is the problem of incomprehensibility
really so serious that it constitutes
a signiicant barrier to access?
At this point — and to complement your
own experience in this matter — I would
like to call as my star witness the former
Swiss Chancellor, Annemarie huber-hotz.
She said the following on this subject in
2000:
‘the greatest challenge in terms of intelligibility
of our laws is currently, without
doubt, EU law, which we have been “independently
shadowing” for a long time
now and, with the advent of bilateral
agreements, directly applying in some
cases. Don’t get me wrong: I am a convinced
European, but, in my view, we
should not just sit back and accept as a
given the sometimes absurdly complicated,
intricate and incomprehensible nature
of some EU laws. As an example, let me
give you only the title of an EU regulation.
I realise that I am expecting a lot of you …
the title is:
“Commission Regulation (EC) No
2592/1999 of 8 December 1999 amending
Regulation (EC) No 1826/1999 amending
Regulation (EC) No 929/1999 imposing
provisional anti-dumping and countervailing
duties on imports of farmed Atlantic
salmon originating in Norway with regard
to certain exporters, imposing provisional
anti-dumping and countervailing duties
on imports of such salmon with regard to
certain exporters, amending Decision
97/634/EC accepting undertakings offered
in connection with the anti-dumping and
anti-subsidies proceedings concerning imports
of such salmon and amending
Council Regulation (EC) No 772/1999 imposing
deinitive anti-dumping and countervailing
duties on imports of such salmon
(Official Journal of the European Communities
L 315, 9.12.1999, p. 17.”’
[Annemarie huber-hotz, Recht haben — gerecht sein
(http://web.archive.org/web/20030701101153/
http://www.admin.ch/ch/d/bk/hu20001106.htm)]
So far, so good — or so bad. however, and
by way of a brief digression, the existence of
possibly incomprehensible EU regulations
prompts an interesting question concerning
EU law. Can an incomprehensible Community
regulation be transposed into national
law so as to be understood, even though it
already applies (but is possibly incomprehensible)
without being transposed?
the Commission, in a recommendation addressed
to Denmark in 1973, considered this
possible:
‘Under Article 189 of the treaty, those
regulations are directly applicable in
member States and entered into force in
Denmark on 1 January 1973. there is
therefore no need to incorporate their provisions
in national instruments. this could
01_2007_5222_txt_ML.indd 40 6-12-2007 15:13:20

ACADEMIC SESSION /
SÉANCE ACADÉMIQUE
lead to confusion as to the entry into force
or the direct applicability of Community
regulations as such. The repetition, even
in part, of provisions of a Community
regulation in a national legal instrument
should at most be permitted only where
necessary in order to make the instrument
comprehensible to users.’
[(73/157/EEC: Commission Recommendation
of 27 April 1973 to the Government of Denmark
on the draft order implementing Regulation No
117/66/EEC and Regulation (EEC) No 1016/68
(Official Journal of the European Communities L 182,
5.7.1973, p. 27). (http://europa.eu.int/smartapi/
cgi/sga_doc?smartapi!celexapi!prod!CELEXnu
mdoc&lg=DE&numdoc=31973h0157&model
=guichett)]
So, there is still surprising scope for national
efforts to make European legislation more
comprehensible.
I am no utopian, and I am not suggesting
that, under current conditions, we could
make all our laws comprehensible to the man
on the street. But should we not aim to
achieve this at least for a core body of law
and certain fundamental principles? If we do
not succeed in credibly communicating this
impression, we may be faced with citizens
mentally leeing the common legal area
which the European Union is trying to build.
After all this, can we say that, having recognised
the problem, we are on the way to
solving it? Given the seriousness of the situation,
this will probably not be enough. But
what more can be done?
firstly, we should never forget what has already
been done by the European Union in
this regard by way of good examples and
decisions. we should mention, above all, the
‘Interinstitutional Agreement of 22 December
1998 on common guidelines for the
quality of drafting of Community legislation’,
the aim of which is to make the Community’s
legal instruments ever more accessible
and comprehensible. this agreement
associates access and comprehensibility in
the context advocated here, and has as its
highest principle:
‘1. Community legislative acts shall be
drafted clearly, simply and precisely.’
[(Official Journal of the European Communities C 73,
17.3.1999, p. 1). (http://europa.eu.int/eur-lex/
lex/LexUriServ/LexUriServ.do?uri=CELEX:3199
9y0317(01):EN:htmL)]
we should also ‘think outside the box’ about
new ways of communicating law. for example,
why not follow the suggestion made by
martin Cutts to start all directives with a citizen’s
summary:
‘few people bother to read directives. One
reason is their strangeness: they are like
nothing most people have read before. Directives
have few features that are welcoming
to ordinary citizens, who may suddenly
ind themselves on unfamiliar ground. A
way of resolving this is to start all directives
with a brief summary of the key points — a
citizen’s summary.’
[http://www.clearest.co.uk/iles/
ClarifyingEurolaw.pdf]
Or, being even more radical, why not try
visual forms of communicating the law? If it
is true that a picture can tell more than a
thousand words, perhaps some of our dificulties
are the result of thinking exclusively
in terms of text.
I would like to conclude with an anecdote
and a question.
I cannot put it better than Jörg Berkemann
did in 1999 at the eighth German EDV-
Gerichtstag (Conference of the German Association
for Computing in the Judiciary) in his
talk ‘free law for free citizens’. the topic referred
to the 18th century, the time of the
Enlightenment, and concerns Johann Georg
Krünitz.
According to Krünitz:
‘National laws need to be brought to the
attention of the general public cheaply by
way of calendars. An annual police calendar
of this kind would, by and by, “bring
to people’s attention the content of old
and prevailing laws, as well as new police
and national laws and regulations, in extract
and through reminders and comments,
thus, over time, bringing whole
collections to their notice.”
In Krünitz’s view, calendars of this kind
were the most effective way of improving
40 | 41
01_2007_5222_txt_ML.indd 41 6-12-2007 15:13:20

knowledge of the law. Adults, children
and servants would read the texts “often,
if only to pass the time” and ultimately
“become fully acquainted with them”. Picture
the scene: the paterfamilias, holding
the latest edition of the calendar, reading
the new legal texts to his assembled family,
checking whether he can remember
what he once knew.’
[Jörg Berkemann, Freies Recht für freie Bürger!, Jur-
PC web-Doc. 188/1999, 1–79. (http://www.jurpc.de/aufsatz/19990188.htm)]
I cannot imagine anything similar catching
on today (if it ever did!).
But what else can we do?
that is still the crux of the matter when it
comes to our understanding of ‘access to the
law’.
01_2007_5222_txt_ML.indd 42 6-12-2007 15:13:20

ACCESS TO EU LAW AND e-LAW —
VISIONS AND CHALLENGES
Royal highness,
Ladies and Gentlemen
we are here to celebrate 25 years of EU law online. this quarter of a century
has not always been an information highway. It has been a long
and winding road — and a success story in providing access to EU law for millions
of users.
As we have noted in the press release concerning this celebration, ‘ignorance
of the law is no excuse’ is still a valid principle. ‘Ignorantia juris non excusat’
is a rule of Roman law and very suitable for EU law online. Other principles of
Roman law, such as ‘in dubio pro reo’, are found in the judgments of the European
Court of Justice in the CELEX and EUR-Lex electronic systems. when the
irst electronic system for EU law got its name, it was a Latin name, Communitatis
Europae LEX, to generate harmony and unity in the multilingual community.
the finnish Broadcasting Corporation has had a weekly news broadcast
in Latin since the late 1980s, showing that the Latin language still prospers.
the finnish EU Presidency supports this tradition with a weekly roundup of
EU news on the Presidency website (www.eu2006.i) in the Latin language.
the old rules and principles are important when discussing the future visions
and challenges of EU law. the principles of Roman law were clear and
logical and easy to implement — in theory. today we live in a different world,
but EU law still needs the same approach — with clear language and logical
legislation. It was agreed back in 1998 in the interinstitutional agreement on
common guidelines for drafting that Community legislative acts shall be drafted
‘clearly, simply and precisely’. I am not sure that we have reached this goal
in all sectors of legislation yet.
It is also an objective at the national level. Last year the European Commission
adopted a communication on better regulation, which aims to improve
the eficiency of the measures for better regulation at both the Community
and the national level, as part of the Lisbon strategy. In finland we have a
national plan of action for better regulation (‘PARSA’, which translates directly
as ‘asparagus’).
KIRSTI RISSANEN
Permanent Secretary,
Ministry of Justice, Finland
42 | 43
01_2007_5222_txt_ML.indd 43 6-12-2007 15:13:21

It is a dificult task to predict the future of EU law, especially EU law online.
I would like to discuss two visions of the future — technical and human.
technology changes, but we humans don’t change. technological progress
has been rapid. As recently as 11 years ago we were not able to use CELEX on
the Internet — I am sure that only a few of you know that CELEX was irst
available via the finnish legal data bank finlex in march 1996! therefore we
can also celebrate today the irst 10 years of CELEX on the Internet.
what will we have in the next 10 years, in 2016? we will probably still be
using EUR-Lex, but we will not necessarily notice that it is EUR-Lex. It will be
a built-in service inside a number of technical tools — all kinds of pods, mobile
toys, word processors and microchips. It is certain that the future of EU law is
e-law — e-law meaning law which is electronic, eficient, ergonomic and
European law. the Official Journal of the European Union will probably be an authentic
electronic journal in 2016, with some paper copies distributed to the
member States.
In 2016, EU law will certainly be more connected to national law. from
the current N-Lex service, giving access to national law in a single portal, we
are moving towards a new legal information network in Europe. this kind of
LINE network gives access not only to law, but also to judgments of national
courts, preparatory acts and legal literature in the member States. E-law will
coexist with e-justice.
what is the human vision? Access to law and transparency in decisionmaking
will be as important as they are today. It is not only getting the texts of
documents but also understanding the contents of the document. Access to
various preparatory documents of EU institutions will be of utmost importance.
the public document registers of the EU institutions serve as a starting
point for further development.
In 2016, citizens and professional users will have easy access to consolidated
texts of legislation. the consolidated, that is updated, texts of original
acts are the most usable texts for all users of legal information — for citizens,
public authorities and the practitioners of law. the original text of any 33-yearold
directive is nothing without the amendments incorporated into it. we already
have access to a large number of consolidated acts today, but by 2016
the coverage will be 100 %.
Understanding is based on language. today we have EU law in 20 languages,
and in two months’ time this will be 23 languages. the multicultural
challenge to all of us is the transformation of the compromised legal language
(some call it Eurospeak) into the language of the citizens. the Austrian phi-
01_2007_5222_txt_ML.indd 44 6-12-2007 15:13:21

ACADEMIC SESSION /
SÉANCE ACADÉMIQUE
losopher Ludwig wittgenstein has stated that: ‘the borders of my language are
the borders of my world.’ how very true. we live in a world which we perceive
according to our cultural framework. we need to remember that ‘no directive
is an island’.
(In finland the courts need to compare different language versions of directives,
in order to ind the ‘correct meaning’ of the law. for the citizen, this kind
of daily analysis is practically impossible. I hope that the online services of EU
law will be of assistance in this problem by providing, for example, summaries
of legislation and a useful ABC of EU law, explaining the terminology.)
Another challenge is the massive low of information and the complexity
of decision-making. Last year a total of 700 new EU regulations, 800 decisions
and 100 directives were published. Even if the number of legislative acts were
smaller in 2016, it would be a challenge to keep pace with this low. therefore
we need some kind of ilter or a messenger, selecting the relevant pieces of
legislation for us. the LexAlert service in EUR-Lex, with user proiles and
e-mail alerts on new legislation, is the irst step in this direction.
(for those who want to know the current phase of any act in the decisionmaking
process, we need a monitoring tool, showing the exact phase in process.
In 2016 we will probably have access to the second-generation version of
the PreLex service.)
As we have noticed, there are huge challenges in the future of EU law. we
can still be optimistic about the future. we have the common European objectives
of better regulation, striving towards simpliication and consolidation of
EU legislation. we also have excellent tools for accessing EU law, with EUR-
Lex, N-Lex and PreLex. ‘where there is a will there is a way’ applies to the future
of EU law online as well. And today we are sure that there is the will.
44 | 45
01_2007_5222_txt_ML.indd 45 6-12-2007 15:13:21

MEETING OF THE
COUNCIL WORKING PARTY ON
LEGAL DATA PROCESSING
24 NOVEMBER
COURT OF AUDITORS
LUXEMBOURG
01_2007_5222_txt_ML.indd 47 6-12-2007 15:13:35

01_2007_5222_txt_ML.indd 48 6-12-2007 15:13:36

SOUVENIRS D’UNE DÉLÉGUÉE NATIONALE AU GROUPE DE TRAVAIL «INFORMATIQUE JURIDIQUE» DU CONSEIL
CELEX: UN SYSTÈME AUTOMATIQUE
DE DOCUMENTATION JURIDIQUE ( 1 )
En tant que déléguée de la Grèce au groupe de travail «Informatique juridique»
et ayant suivi régulièrement les travaux de ce groupe au sein du
Conseil de ministres, je me réjouis d’avoir la possibilité de fêter les vingt-cinq
ans du droit communautaire en ligne et plus de vingt-cinq ans d’existence de
ce groupe et, surtout, de son travail fructueux et eficace.
C’était en 1983 que mon pays, la Grèce, se réjouissait du nouveau système
Celex, prometteur et qui nous avait éblouis. J’écrivais alors à l’époque, dans
une revue juridique grecque ( 2 ), combien il était nécessaire de disposer d’un
système donnant accès à toute l’information juridique et judiciaire, tant pour
les services publics et les universités que pour les particuliers qui s’intéressent
au droit et à son application. Cette nécessité de recherche documentée est devenue
encore plus aiguë en droit communautaire vu le nombre d’actes adoptés,
la diversité des secteurs couverts, la nature spéciale de l’ordre juridique de
chaque état membre, ainsi que la diversité des langues oficielles.
Par sa résolution du 26 novembre 1974 concernant la documentation juridique
automatique, le Conseil des ministres de la Justice avait lancé l’idée de la
création d’un système communautaire de documentation juridique automatique
ouvert à tous les états membres, puisque ceux-ci devaient appliquer le
droit communautaire.
Dans l’article, je me référais aussi au champ couvert par le système, qui
concernait non seulement les textes des traités, mais également le droit dérivé,
les accords conclus par l’Union européenne (à l’époque la Communauté européenne),
la jurisprudence de la Cour de justice ou encore les actes préparatoires.
Ce qui était étonnant, c’était que le système s’étendait aussi aux actes adoptés
par les états membres pour mettre en œuvre le droit communautaire au sein de
leur ordre juridique interne et mentionnait déjà la jurisprudence nationale. Le
GALATEIA ALEXAKI
Déléguée nationale grecque
au groupe de travail
«Informatique juridique»
du Conseil
( 1 ) m me G. Alexaki n’est pas intervenue dans le cadre de la réunion même, mais a eu la
gentillesse de nous envoyer un texte relatant ses souvenirs du début de la participation
de la Grèce au groupe de travail «Informatique juridique» du Conseil.
( 2 ) Alexaki, G., «to koinotikó sýstima autómatis nomikís tekmiriósis (Celex)», Nomikó Bíma,
1983, p.1339.
48 | 49
01_2007_5222_txt_ML.indd 49 6-12-2007 15:13:36

système incluait des actes depuis 1952 et la mise à jour se faisait avec un délai
de trois à quatre semaines à partir de la publication de l’acte. Chaque institution
communautaire était chargée des secteurs la concernant. Le Conseil fournissait
ainsi des données concernant les traités, les relations extérieures, le droit dérivé,
le droit complémentaire. Le Parlement européen, la Commission et le Comité
économique et social fournissaient les actes préparatoires. La Cour fournissait
la jurisprudence et le Parlement les questions parlementaires.
La procédure de chargement débutait par un formulaire en deux parties.
Une partie, assez détaillée, comprenait plus de trente rubriques de données
documentaires, et une seconde présentait le texte entier ou un résumé de celuici,
selon le schéma du formulaire. On pourrait mentionner certains types de
rubriques, telles que l’institution émettant l’acte, le titre de l’acte, sa date de
publication, sa date d’entrée en vigueur, la date de sa transposition dans l’ordre
juridique interne, la date du débat au Parlement européen, l’incidence sur
d’autres actes (abrogation, modiication, etc.).
La recherche dans le système CELEX se faisait sous forme de questions
posées par l’usager en utilisant des mots clés ou des codages. Le résultat se
présentait à l’écran et pouvait également être imprimé.
L’évolution du système CELEX, plus largement mis à la disposition du public
à partir du début des années 1990, avait débuté par un premier plan de développement
lancé par le groupe de travail du Conseil en janvier 1978 pour la
période 1978-1982. Ce plan prévoyait la création des secteurs non encore existants,
l’accès dans d’autres langues oficielles pour l’ensemble des institutions
communautaires, ainsi que l’ouverture, à tous les états membres au niveau de
leurs administrations, de l’ensemble des secteurs couverts par le système.
Les résultats n’étaient cependant pas encore satisfaisants à la in de 1982, et
cela était dû tant au changement du software et du hardware du système qu’à des
ressources en personnel insufisantes. malgré tout, le système a progressé: alors
qu’en 1981 seule la version française existait, les versions anglaise et allemande
ont été ajoutées en 1982. Un nouveau plan d’action fut alors adopté pour les
années 1983-1985 avec notamment les actions suivantes: une étude de faisabilité
pour la langue grecque, combinée avec une phase expérimentale en 1985.
La réalisation de ce plan exigeait l’existence de ressources sufisantes.
Je terminais alors mon article dans la revue Nomikó Bíma en soulignant que
le système CELEX constituait un moyen important pour le service du public et
surtout les milieux juridiques et judiciaires qui souhaitaient connaître le droit
communautaire. J’ajoutais qu’il était souhaitable de promouvoir en Grèce
l’évolution de l’informatique juridique ain que de plus en plus d’administrations,
d’organismes et de services puissent être raccordés au système CELEX.
01_2007_5222_txt_ML.indd 50 6-12-2007 15:13:36

EUR-LEX TODAY AND TOMORROW
After more historical and philosophical speeches on the development of
access to law, speaking of what has been achieved recently in EUR-Lex
and what is planned for 2007 is a bit like coming down to earth. But managing
a database on a daily basis is down-to-earth work and it is of course a pity that
we do not have more time to dream about the perfect access to European law
or, more pragmatically, to make more high-lying plans for the evolution of
EUR-Lex.
where do we stand in November 2006? two years after the launch of the
new EUR-Lex on 1 November 2004, the system has been fully stabilised in
2006: users — who visit all corners and possible combinations of the system
to a greater extent than we do — have helped us to ind bugs and get rid of
them. EUR-Lex is working in a stable environment based on a stable infrastructure
which lets us cope with increasing numbers of users. the full documentary
structures have been safely kept: EUR-Lex is not only the access to
documents, it is also the access to legal information, crosslinking documents
and information. the documentary structure allows the constitution of networks
of documents and of legal information.
Different parts of the system already deal with the 23 languages which
will be oficial EU languages from 1 January 2007: the provisional site, with
provisional versions of documents in Bulgarian and Romanian, has been available
since June 2006, the treaties are available in Irish and former editions of
the Oficial Journal published in Irish since 1998 have been loaded and made
available.
By mid-November 2006, the number of users was around 200 000 on
working days, and was increasing.
the system still needs further development and the EUR-Lex team still has
new ideas to implement. Some of these developments will be immediately
obvious to the user, but there are also behind-the-scenes elements which are
important for an appropriate management and handling of the back ofice of
the database.
PASCALE BERTELOOT
Head of the EUR-Lex unit,
Publications Office
50 | 51
01_2007_5222_txt_ML.indd 51 6-12-2007 15:13:38

1. hOw DO wE END 2006?
Expert search was made available in August and the advanced search link
giving easier access to expert search has now been activitated. Expert search is
not a completely new development as it was in fact cut out of CELEX and had
to be reconnected with all its ‘capillarities’ to the new EUR-Lex system. Expert
Search is a very potent tool, essential for specialists, in particular within the EU
institutions and in national administrations.
LexAlert, the system notifying new documents or information to users,
has been in production for several months even if registration online is still not
possible. the weak point of notiication systems is the e-mail infrastructure:
the risk of the sending server being identiied as an author of spam is quite
high.
that is why notiication is also offered by RSS feeds: this functionality
should develop in the future and might become the only one available for
mass notiication.
Besides linking and RSS feeds, new features for 2006 related to the reuse
of data. web services have been developed and tested internally within the
institutions. they allow, for example, speciic systems maintained by directorates-general
of the Commission to include data from EUR-Lex as the reference
database. Duplication of documents and information can thus be avoided and
updating in parallel to EUR-Lex also becomes possible. web services should
also interest member States, for example in a database following their own
implementation of directives.
traditionally, CELEX and now EUR-Lex are licensed. New formats and
the new EUR-Lex allowed for new formats for licence holders, to which we
added a new communication mode. An ftP store server for the licence holders
was opened by mid-November and the irst extractions in XmL delivered.
Since summer 2006, one main task has been preparing for enlargement.
the special editions of the Official Journal of the European Union which publish
oficially all acts in force on the day of accession in the languages of the new
member States have also had to be available online in Bulgarian and Romanian.
In addition the current editions of the Oficial Journal will have to be made
available in the new languages from 1 January 2007. Developments are well
under way and the results should be available in December 2006. further developments
are needed for creating all browsing and search functions in the
new languages and uploading documents in the new languages as soon as
they are made available.
01_2007_5222_txt_ML.indd 52 6-12-2007 15:13:39

MEETING OF THE COUNCIL WORKING PARTY
ON LEGAL DATA PROCESSING
One current task is related to improving the quality of information on
national implementation measures: it requires close collaboration with the
Secretariat-General of the Commission. the Secretariat-General’s new database
can now systematically deliver for dissemination references to the different
measures adopted in each member State. the information on national
implementation measures is part of the bibliographic notice of directives. further
structuring of the information in the future would make it easier to link to
the full text of measures in national legislative databases which offer a URN or
URI server.
the last development offered to the user before the end of 2006 will be the
access to judgments of the Civil Service tribunal set up by the end of 2005.
New editorial features include a simple search guide and an advanced
search guide online.
2. mAIN DEVELOPmENtS fOR 2007
for 2007, there are four main themes: enlargement and a new linguistic
regime, enrichment of document collections, further developments of the system,
and usability. Besides these main tasks, the Publications Ofice will also
work on a few other elements to facilitate legal information.
Enlargement and a new linguistic regime
Enlargement has already been mentioned, but in addition Irish will become
a full oficial language — although with a derogation similar to the
one adopted for maltese — from 1 January 2007. Dealing with the Oficial
Journal is not all. Step by step all functionalities in EUR-Lex will have to
include three new languages and offer the users of these new languages the
same service offered in the older linguistic versions. this will not mean that
all documents will be available in the new languages, but that information
will at least be available whatever linguistic version of EUR-Lex is used.
Browsing access to case-law will, for example, be needed from the beginning
of January as documents of the European Courts will be available in
Bulgarian and Romanian on the day new judgments are given. the same
applies to COm documents, which should also be delivered in the two new
languages from the beginning of 2007. As for the acquis communautaire, the
expectation is to terminate publishing by summer 2007 and then start the
uploading of iles in the new languages. (there will be no acquis translated
into Irish.)
52 | 53
01_2007_5222_txt_ML.indd 53 6-12-2007 15:13:39

Enrichment of collections
the enrichment of collections is the result of requests from users. tables
and annexes of legislative acts are available in PDf if they were published after
1998, but they are not available in htmL. On the basis of the versions available
in XmL since 2004, we should be able to offer tables and annexes of legislative
acts in 2007.
Images also appear in court decisions, speciically in intellectual property
and trademark cases: not seeing the image impedes full understanding of the
case. EUR-Lex will process images in documents of the European Courts as
from 2007.
Although the Oficial Journal has been available online since 1998, the OJ
tables have not been used in EUR-Lex. they are quite an interesting instrument
on paper, but an even more interesting instrument if used electronically.
they offer information and possibilities to cross information and make documents
available from another point of view.
Press releases related to documents of the European Court should be included
in the LexAlert notiication system. Even if they will not be available
through EUR-Lex search functionalities, they can be assimilated to a way of
disseminating legal information that is closer to the citizen, just as a future
summary of legislative acts for citizens could be for legislation. Just note that
the Court does not issue press releases for all judgments or opinions: they refer
only to the most important judgments.
Consolidated legislative texts have been made available in PDf through
EUR-Lex, but they have never been loaded in htmL. the whole collection is
available in XmL format in-house and using it in EUR-Lex would give a new
dimension to the dissemination of consolidated legislative texts which represent
Sector 0 in EUR-Lex: they would be indexed and also belong to extractions
for licence holders.
Further developments
As to further developments in 2007, only three elements will be mentioned
here.
the longer-term intention is to give early access to all acts belonging to the
legislative procedure. EUR-Lex already gives access to COm documents. these
are mostly Commission legislative initiatives, but they are now often accompagnied
by SEC documents which contain impact assessments: the main role
of those documents is to demonstrate the implementation of the Lisbon strat-
01_2007_5222_txt_ML.indd 54 6-12-2007 15:13:39

MEETING OF THE COUNCIL WORKING PARTY
ON LEGAL DATA PROCESSING
egy. formerly these documents were never sent to the Publications Ofice and
even now the full collection will not be made available on EUR-Lex. the aim
remains to give more visibility to the impact assessments among acts of the
legislative procedure in EUR-Lex.
Since the Court of Justice irst started to put its documents into CELEX in
the 1970s, provision was also made for a sector giving access to information
on national case-law which the Court had been asked to collect by a Council
resolution of 1978. the irst phase of development will include the loading of
the data available within the Court itself, and provision for a continuous uploading
procedure. Phase 2 will include the development of appropriate search
functionalities. the data collected by the Court and the documentary work
on these data does not include the full text of judgments but information —
including under more descriptors — and an analysis. for the past — and the
oldest record dates back to 1959 — making the full text available is a copyright
issue. for newer cases, linking to stable national databases is of course
an option.
Until now, one of the options for browsing — parliamentary questions —
has remained undeveloped. EUR-Lex offers a link to the European Parliament
site, but this does not insert the parliamentary questions in the context of
other legal documents to which they are related. If looking for a certain theme,
a user might miss important information or views provided by a parliamentary
question and the answer to it. formerly, parliamentary questions were published
in the Oficial Journal, but very often quite a long time had elapsed since
they had been answered. the Publications Ofice is in contact with the Parliament
on the best way to give early access to questions and answers and constitute
the right metadata set.
Usability
the main rules followed when designing the EUR-Lex interfaces were
clarity and consistency. the Publications Ofice nevertheless ordered a usability
study on EUR-Lex. the study included tests with users regarded as nonprofessional,
i.e. neither lawyers nor documentalists.
Usability is generally deined by three criteria: the time a user needs to use
a system, the time he or she needs to execute tasks and the number of errors
he or she makes before getting adequate results. the inal report of the study
is expected to be available in february 2007. the outcome will then be
analysed and ways of implementation examined. for the time being, the
54 | 55
01_2007_5222_txt_ML.indd 55 6-12-2007 15:13:39

Publications Ofice estimates that three categories of actions will be required: easy
elements which can be implemented right away, medium-term elements which
require some development, and long-term elements which might include redesigning
the information structure or deeper relection before implementation.
this last category might imply the ordering of a more detailed study. Besides
the usability study, suggestions or requests for adaptations or changes
might come from users with whom the Publications Ofice is in contact —
mainly through the help desk — from the Groupe interinstitutionnel Lex (GIL),
the Council working party, and from professionals encountered from time to
time at conferences, fairs, etc.
Further facilitating the access to legal information
the ‘Directory of Community legislation in force’ has its origins in a request
of the Council in the 1970s. Since the new EUR-Lex was created, the
directory has been updated monthly. Its structure allows for a stable classiication
of acts. within the database, preparatory acts are also classiied by the
chapters and sub-chapters of the directory. this means that the available metadata
allow for a directory of preparatory acts belonging to still-pending legislative
procedures, which could also be updated monthly. A speciic directory of
consolidated versions could present only the last available layer of consolidation,
mentioning — if this were the case — the last modiier that had not yet
been integrated.
the last task which will be mentioned here is PreLex. By a resolution of
19 December 2002, the European Parliament required a better integration of
various information systems existing in the ield of law and mentioned under
more (old) EUR-Lex, CELEX and PreLex. Old EUR-Lex and CELEX have now
become a new integrated system. for PreLex, the Secretariat-General of the
Commission decided that its core business certainly included the production
of the data and the use of them internally but that the dissemination falls under
the competencies of the Publications Ofice. Negotiations about the handover
of the dissemination of PreLex data are under way and by the end of 2007 the
public system of PreLex should be with the Publications Ofice. In addition to
ensuring regular production of the dissemination data and maintenance of the
system, the Publications Ofice will also begin to analyse how the PreLex data
could be integrated into the existing sets of EUR-Lex metadata in order to
merge EUR-Lex and PreLex, knowing that the same information on legislative
procedures as today should continue to be available to the user. this would
01_2007_5222_txt_ML.indd 56 6-12-2007 15:13:39

MEETING OF THE COUNCIL WORKING PARTY
ON LEGAL DATA PROCESSING
mean that in the long term PreLex information would be based on further enriched
metadata in EUR-Lex and merged with EUR-Lex.
3. thE fUtURE BEyOND 2007
managing a database is down to earth, but it should not prevent us from
seing EUR-Lex in perspective. the president of the academic session yesterday
mentioned 2016. Let’s however stick with 2010 or 2012. EUR-Lex will have
undergone an evolution and hopefully the right evolution, but could we also
develop a vision like the visions of the 1960s which gradually created CELEX
and EUR-Lex? what suggestions can be put forward?
One proposal is to establish better connections to research in the ield of
legal informatics. Of course, research in legal informatics has different levels:
irst there are steps towards a solution which might become reality after many
years, then research which brings concrete results in the medium term, and,
inally, research which delivers conceptual solutions which could be analysed
for implementation in the short term. to better connect to research in the ield
of legal informatics, the Publications Ofice could promote the use of EUR-Lex
content for research purposes.
EUR-Lex content can be made available for researchers under three conditions:
the research project has to have a multilingual dimension, the content
cannot be used commercially, and the Publications Ofice is entitled to receive
the various reports on the research results. Another means is also to follow
research projects which are published or presented at conferences in order to
follow the long-term evolution of thoughts and horizons in the ield.
Let us just take a few examples.
the concepts of ontologies, taxonomies and topic maps have been circulating
in the research community on legal informatics for many years. Prototypes
have been developed, but mostly with very narrow contents. how could
these concepts be applied to such multilingual corpora as EUR-Lex? If we admit
that it is not possible to fully create new documentary tools ‘manually’ and
apply them to hundreds of thousands of units, it might be possible, on the
basis of existing and used documentary tools, to reorganise the information
retrieval.
Online documents are still presented in the way in which printed versions
would appear. Should totally different rules be followed or should the length
of lines, at least, be deined morphologically, to make them more readable for
users?
56 | 57
01_2007_5222_txt_ML.indd 57 6-12-2007 15:13:39

Complexity is a major theme in various ields, including the representation
of complex systems by computer systems. Do documents and metadata in
EUR-Lex correspond to the deinition of complex systems? how should legal
complexity be represented? how can legal information be modelled? Is not
one of the challenges of legal information systems to present legal complexity
in an understandable way? to recognise the complexity of law is not an action
against simpliication, which indeed is necessary in all its features (consolidation,
codiication, reducing the administrative burden, etc.). the complexity of
law is basically linked to the complexity of life and new factors making it even
more complex appear every day! we didn’t need so many regulations on
media 30 years ago!
making models of legal information can also help in understanding the
evolution of the law. One of the ways of modelling lies in the paradigm of
networks.
A project analysing 26 000 decisions of the Supreme Court of the United
States and mapping them in the network in which they exist through their
mutual quotation shows which were the most important concerns of the
Court independently of the issue of the case itself. It showed, for example, that
after the Second world war the most important decisions and the most important
quotations were on freedom of speech. Such models give totally objective
information on the content of court decisions, free from the human appreciation
of analysts and academics.
this means, inally, that the EUR-Lex content is not only at the disposal of
researchers to contribute to the long-term evolution of the information system,
but also for various interpretations. Legal research could discover a new dimension.
01_2007_5222_txt_ML.indd 58 6-12-2007 15:13:39

DOCUMENT ANALYSIS
AND LEGAL INFORMATION
Speaking in the framework of celebrating ‘25 years of EU law online’ offers
a most welcome opportunity to highlight some aspects of the daily work
of the EUR-Lex team that one might consider of some importance, in particular
for the end-users of EUR-Lex.
first, some emphasis will be laid on the system’s capacity to produce, in
response to a query, a well-deined set of documents, which is to be distinguished
from the rather simple search for only one particular reference.
Second, the multilingual nature of the system will be further elaborated
on. Presenting how the Eurovoc thesaurus and the classiication of the ‘Directory
of Community legislation in force’ can support the user in bridging the
shortcomings that a search for a natural language term in the full text of a
document can provide is of even more interest with regard to the multilingual
context of EUR-Lex.
MICHAEL DÜRO
Documentalist
in the EUR-Lex unit
of the Publications Office
1. SEARCh OPtIONS: A RESULtS PERSPECtIVE
It is, on this occasion, hardly possible to deliver the full picture of EUR-
Lex’s simple and advanced search options. the perspective chosen for this
presentation rather is to focus on two aspects by introducing two practical
examples.
1.1. Search options for a single document
the irst sample query illustrates the search for one, and only one, document.
the search for one single document is most conveniently performed by
using a unique identiier as search criterion. As such a means of unique identiication,
the EUR-Lex simple search offers the searches by document number
or by publication reference.
the heading ‘Search by document number’ covers the ‘natural number’ of
documents, e.g. a regulation or a COm document, together with the year. the
same criteria can be applied to a search for ‘consolidated texts’ under the same
heading.
58 | 59
01_2007_5222_txt_ML.indd 59 6-12-2007 15:13:39

the CELEX number is, in its default format, made
up of information concerning the type of document and
the collection it belongs to, its year and natural number,
and serves as a unique identiier across all sectors and
document types within the EUR-Lex system.
the heading ‘Search by publication reference’ also
allows for the narrowing down of a query to retrieve
only one single document, but is limited to the documents
published in the Official Journal of the European Union
or the European Court of Justice’s Court Reports. the details needed to
perform such a search by publication reference include, for the Oficial Journal,
the year, the OJ series (L for ‘Legislation’ or C for ‘Information and Notices’) as
well as the number of the irst page of the document and inally either the exact
date of the OJ or its number.
for the European Court Reports, the year and the page number together
with a selection of the volume (Court of Justice, Court of first Instance, Staff
Cases) are necessary to limit the result to a single document.
1.2. Search options for a set of documents
the most simple and, one might wrongly assume, maybe least interesting
solution to retrieve a well-deined set of documents in EUR-Lex is to combine
several simple searches for a single document, e.g. by CELEX number. however,
this approach comes with the big beneit of being very convenient to the
user.
the irst practical example introduces three individual queries for codiication
projects.
‘Codiication is the process that brings together the provisions of existing
acts with all of their subsequent amendments into one law’ (cf. COm(2006)
689, p. 7) and must not be confused with consolidation, the result of which
only serves information purposes and is not legally binding.
the starting point for the codiication process is a draft for a codiied version,
which is provided for by the Commission, and which represents a new
document type, which recently had to be introduced into the collections covered
by EUR-Lex. the Commission aims at inishing the work on a total of
500 such codiication projects up until 2008, out of which a test set has already
been delivered to the Publications Ofice to be loaded into the EUR-Lex
database.
01_2007_5222_txt_ML.indd 60 6-12-2007 15:13:40

MEETING OF THE COUNCIL WORKING PARTY
ON LEGAL DATA PROCESSING
1.2.1. SEARCHING FOR A SET OF DOCUMENTS:
COMBINING SINGLE DOCUMENT SEARCHES
As a irst step in our example, three individual searches by CELEX number
(DN) are performed to retrieve results for the overall number of codiication
projects available in the database concerning:
• regulations (DN=5????EC*),
• directives (DN=5????GC*) and
• decisions (DN=5????fC*).
the three result lists can be combined by Boolean
operators (AND, NOt, OR) in the ‘Search history’ section
within the simple search.
At the time of writing this article, these queries result
in only one hit each. Consequently the combining of the
results with a Boolean ‘OR’ produces a list of three documents.
the exploitation of this particular feature of the ‘Search history’ extends
the simple search’s capacity and allows for the easy combination of multiple
simple queries to simulate and replace, to a certain extent, comparably complex
searches.
Achieving the same results might sometimes turn out much more dificult,
when trying to produce them through some repeated application of the ‘Reine’
option coming with every result list. however, both of these tracks simulating
more complex searches within the simple search environment are limited
to the ields in the EUR-Lex database that can be addressed via the simple
search screens.
But EUR-Lex has more to offer than the simple search, and goes, concerning
the functions available, far beyond the simple search for documents.
to learn more about that, and if it comes to exploiting further ields of the
database, one needs to turn to the advanced search: almost all ields available in
EUR-Lex are available for selection from a list in the advanced search screen.
1.2.2. SEARCHING FOR A SET OF DOCUMENTS:
COMBINING ANY FIELDS
the second practical example was chosen to illustrate the easy combination
of search criteria, and is derived from a question that was recently addressed
to the EUR-Lex unit of the Publications Ofice:
for how many documents adopted one year before, respectively after, the
2004 enlargement, did the Council need to act unanimously?
60 | 61
01_2007_5222_txt_ML.indd 61 6-12-2007 15:13:40

Some treaty articles, which require the Council to
act unanimously when adopting an act, were used as
main search criteria and applied to the ield indicating
the legal basis of an act. the time frame concerning the
date of the document and some limitations to certain
collections (Sector 2: international agreements; Sector 3:
secondary legislation; Sector 4: supplementary legislation)
were translated into additional search criteria.
the resulting query in the advanced search, which produced 25 hits,
looked like this:
[Date_of_document > = 2003/05/01 < = 2004/04/30 AND
(Legal_basis = 12002E018 OR Legal_basis = 12002E042 OR
Legal_basis = 12002E047?P2 OR Legal_basis =
12002E151P5) AND (Type_Sector = 2 OR Type_Sector = 3
OR Type_Sector = 4)]
the legal basis ield, which is of central importance when performing this
example query, is not available from the simple search screen.
to get a comparable result for the one-year period after the enlargement of
1 may 2004 and consequently with 25 instead of 15 member States being
represented in the Council, one only has to adapt the period covered in the
‘date of document’ part of the query. for the irst year after the date of enlargement
the query results in seven hits; for the period 1 may 2005 to 30 April
2006 nine references are retrieved.
for further information, and when looking for a general introduction to the
advanced search, it is recommended to consult the ‘Visual quick start guide: advanced
search’, which is available online in the EUR-Lex help section (http://eurlex.europa.eu/en/tools/help_advanced.pdf)
(last visit 24.11.2006).
this irst part of the presentation aimed at fuelling the user’s imagination
when it comes to searching not only for individual references, but for welldeined
sets of documents, and exploiting the more than 120 ields available in
the advanced search for that purpose.
2. EXPLOItING EUR-LEX AS A mULtILINGUAL DAtABASE
the second part of the presentation will elaborate further on some of the
multilingual aspects of EUR-Lex and focuses on the Eurovoc thesaurus and the
classiication of the ‘Directory of Community legislation in force’.
01_2007_5222_txt_ML.indd 62 6-12-2007 15:13:41

MEETING OF THE COUNCIL WORKING PARTY
ON LEGAL DATA PROCESSING
Both tools aim at supporting the user and both tackle certain language issues
related to the database.
2.1. The Eurovoc thesaurus
the irst edition of the Eurovoc thesaurus dates back to 1984. today this
controlled vocabulary is available in version 4.2 in 18 oficial languages (exceptions
are maltese and Estonian) and comes with approximately 6 400 descriptors,
an almost equivalent number of reciprocal hierarchical relationships (broader–
narrower term) and around 3 500 reciprocal associative relations (related term).
the added value that this tool provides for EUR-Lex users cannot immediately
be derived from these purely statistical igures. But queries based on natural
language search terms already come with certain weaknesses when applied
to a monolingual database. Cases of homonymy and synonymy can be referred
to as blurring the results of that kind of query. Nevertheless, quite a few users
tend to perform this kind of search irst — and sometimes only this kind.
the natural language search term ‘electricity’ was
found in the title of 573 documents (EN advanced search
query [tI=electricity]). for comparison, in the Italian language
version of the database, the search in the title for
the synonyms ‘corrente elletrica’ AND ‘elletricità’, combined
with a Boolean OR, produced only 278 hits (It
advanced search query [tI=corrente elletrica OR
tI=elletricità]) and applying ‘Strom’ OR ‘Elektrizität’ in
the German language (DE advanced search query
[tI=Strom OR tI=Elektrizität]) gathers only 149 references. One reason for the
comparably low number of hits in the German example is the availability of
numerous composites in the German language which are not retrieved by that
example of a query.
Consulting the Eurovoc thesaurus results, for all ive terms that were used
for the above queries, in only one common descriptor, which is ‘electrical energy’
in English, ‘energia elettrica’ in Italian and the corresponding ‘elektrische
Energie’ in German. the terms previously used for the search in the title all
serve as non-descriptors in the controlled vocabulary and provide a reference
to the descriptor.
Applying the descriptor for a search exploiting the Eurovoc ield in the
EUR-Lex database (e.g. [DC=electrical energy]) results, for all the three language
versions mentioned, in the same number of hits (which was 1 123).
62 | 63
01_2007_5222_txt_ML.indd 63 6-12-2007 15:13:42

"
Exploting a
multilingual database
electricity
Natural language search term: TI=
corrente elettrica
OR elettricità
Strom
OR Elektrizität
573 278 149
1123 1123 1123
electrical
energia
elektrische
energy
elettrica
Energie
EUROVOC descriptor: DC=
the results were not only consistent across the language versions; they also
seemed to be more complete, as the user also avoided the risk of missing
important documents for which the chosen search term simply does not appear
in the title.
Another beneit of the application of the Eurovoc thesaurus, which only at
the irst glimpse appears to be of minor importance to the EUR-Lex user, is that
the intellectual analysis of the content of the document, which is the basis for the
attribution of the descriptors, can be performed independently of the language
that a future user might select for the consultation of the thesaurus when searching
for documents. Consequently the effort spent for that part of the analysis
multiplies per language version in which the thesaurus is available.
the Eurovoc thesaurus aims at overcoming the challenges that EUR-Lex
potentially provides as a multilingual system, by supporting language coverage
as completely as possible for all oficial languages. A different approach is offered
by the classiication of the ‘Directory of Community legislation in force’.
2.2. The classiication of the ‘Directory of Community
legislation in force’
this tool is used to index legislative acts (including the preparatory works)
and provides a numerical classiication system, which comes with 20 headings
at entry level. Each of those is marked by two digits (from 01 to 20). the ‘15’,
for example, represents ‘Environment, consumers and health protection’. Below
the entry level up to three more levels are possible. Overall the classiication
comprises an identical set of 465 (sub-)headings in all — at the time — 20
oficial languages. the numerical structure as such guarantees language independence.
01_2007_5222_txt_ML.indd 64 6-12-2007 15:13:42

MEETING OF THE COUNCIL WORKING PARTY
ON LEGAL DATA PROCESSING
But its hierarchical set-up also provides for convenient browsing and navigation
through the different levels of headings and the underlying subheadings.
Again, the advanced search adds a convenient feature: the selected
application of truncation symbols allows for the exploitation of the directory
code to search for horizontal criteria relected in the classiication.
On the second hierarchy level, i.e. in the numerical system represented by
digits three and four, the combination ‘07’ represents the horizontal criterion
of ‘statistics’. Consequently, under each of the 20 headings that provide acts
on statistics, those references are gathered, inter alia, under this particular subheading.
the query [CC=0?07+ OR CC=1?07+ OR CC=2007+] produces a list of
all documents (in Sectors 2, 3, 4, and 5) presumably dealing with statistics that
are available in the database. for the time being the query resulted in 1 269
hits, the retrieving of which would be almost impossible without the exploitation
of the classiication of the ‘Directory of Community legislation in force’.
SUmmARy AND CONCLUSION
this presentation aimed irst at giving an overview of the EUR-Lex search
options allowing the retrieval of one single document. the next step, introducing
a irst practical example, was to combine the results of three queries using
the CELEX number via the search history of the simple search to produce a
well-deined set of documents as a result. Leaving the simple search and moving
on to the advanced search then allowed for the combination
of any of the ields available in the database.
the second example came with a comparably complex
query and exploited a ield that cannot be addressed
by any simple search option (e.g. the ield containing information
on the legal basis). for those further interested
in exploiting the additional functions and various search
options the advanced search offers, the ‘Visual quick
start guide: advanced search’ is recommended reading.
In the second part of the presentation, examples were used to demonstrate
how the Eurovoc thesaurus and the classiication of the ‘Directory of Community
legislation in force’ can support the user in retrieving better results and
going beyond the rather weak searching for natural language search terms.
Due to the multilingual character of EUR-Lex, the two instruments offer even
more value to the system.
64 | 65
01_2007_5222_txt_ML.indd 65 6-12-2007 15:13:42

EUR-Lex seems suitable to serve the occasional user well when addressing
simple queries to the system. with regard to achieving optimised recall and
precision values, EUR-Lex offers powerful tools, some of which play an even
more important role in the multilingual context. making the effort to get more
familiar with some of the more elaborate options is, with regard to the results,
worth it, and not only for the professional or expert user.
01_2007_5222_txt_ML.indd 66 6-12-2007 15:13:42

LIFE AS A CELEX HOST
INtRODUCtION
It is a real pleasure to be here today and to celebrate the anniversary of
CELEX.
I have been asked to say a few words about life as a CELEX host. first, I
will say something about why we became a host and why we still are a host
16 years later. Second, I will say something about the problems we have encountered
during this time.
Lovdata’s life as a CELEX host started in summer 1990. At that time Norway
was preparing for its second application to the European Communities,
and the ministry of foreign Affairs was publishing a series of position papers
on various aspects of the common market. Since these papers were part of the
Lovdata system and contained a lot of references to EC directives and regulations,
we thought it would be nice to be able to directly link to these.
TRYGVE HARVOLD
Director, Lovdata, Norway
GENERAL ImPRESSIONS
Before I continue with the more speciic experiences of our new life as a
host, I would like to share with you some general impressions from these
years.
thE PUBLICAtIONS OffICE
first of all, the experience has been very agreeable overall. As you all know,
the Publications Ofice has always been professional and competent, providing
detailed documentation for the data and delivering new material in a timely
fashion. Of course, there have been a few hiccups in the operation, but these
have, more often than not, been due to our own shortcomings.
thE CELEX DOCUmENt NUmBER
Perhaps the thing that impressed us the most, when we got acquainted
with CELEX documents, was the CELEX document number. Everyone who
66 | 67
01_2007_5222_txt_ML.indd 67 6-12-2007 15:13:44

has worked with legal databases knows that it is not easy to establish links
based on citations in the text, because for many types of documents there are
no standard rules for the citation format, or there are some rules, but these
have changed over the years. for us it was a revelation to discover that CELEX
used a general format for all document types, and that the format was used
systematically both as document identiier and in citations. In addition, a
CELEX document would refer to other documents which cited the document
in a signiicant context. the CELEX number also said something about the
document itself — besides identifying the year and number, it said something
about what type the document was. we did ind it a little curious that the year
of the document was only represented by two digits, implying perhaps that at
some time somebody had not been totally convinced that the European Communities
was a project for the long-term future.
thE REASONS fOR BEING A hOSt
Several people have asked me why we still bother to be a host when all the
CELEX sectors are available free of charge in EUR-Lex. this may seem all the
more strange since EUR-Lex is a truly excellent system which is easy to navigate
and search.
my response is that by integrating CELEX and our national system we
achieved easier access to CELEX documents overall by Lovdata users. more
speciically we have established:
• chronological tables of content for all CELEX sectors,
• links from all national documents which cite CELEX documents, that is
from the national documents where the regulations are incorporated in
Norwegian law,
• links from all CELEX documents to national documents which cite CELEX
documents, so that if a national document cites a CELEX document, there
is established a link from the CELEX document to the citing document,
• easy and fast searching of CELEX documents on a par with our other databases.
Links are very important in our system. By establishing these links, which
are used very actively, I think we have obtained our aim of integrating Community
and national law at least to a certain extent.
historically there has been one other important reason for wishing to integrate
CELEX with our national system. As you know, Norway is not part of
the European Union, but together with the other EftA states we are part of the
01_2007_5222_txt_ML.indd 68 6-12-2007 15:13:44

MEETING OF THE COUNCIL WORKING PARTY
ON LEGAL DATA PROCESSING
European Economic Area (EEA), which means that many directives, regulations
and decisions apply just as much in Norway as in the rest of Europe. the
directives and regulations that apply in Norway must be translated into Norwegian
by the ministry of foreign Affairs, and we have built a Norwegianlanguage
version of CELEX based on these documents. So our CELEX contains
English and Norwegian documents, as well as the Danish version of Sector 6
(court decisions), since EC Court decisions are not translated into Norwegian.
we established our irst CELEX system in 1990 as part of our online system.
In practice this meant that all the CELEX documents were automatically
given our structural codes, which were a subset of the SGmL family. the coding
was not very detailed, because there was not much structural detail in the
CELEX documents to begin with. But it did mean that when we established
our irst home page on the web in January 1995, we could start considering
putting our English version of CELEX up on the web. In summer 1996 we
demonstrated our CELEX on the web for the Publications Ofice in Luxembourg,
and made a general demonstration at London Online in December of
the same year. In the following years we marketed this version as a separate
system, mainly in the Nordic countries, although we had subscribers all over
the world.
mEmORIES Of tEChNICAL CURIOSItIES AND mIShAPS
when you are a host for more than 15 years, some curious things happen
along the way. Some of those things have their roots in the past. the It world
looked very, very different 25 years ago, and I think it is amazing that CELEX
has actually been in operation that long. however, these deep roots into the
past do throw up a few unexpected technological curiosities.
thE tAPES
when we irst applied for the English version of CELEX, we received a
stack of industrial nine-track tapes. the use of nine-track industrial tapes
brought back fond memories for those of us who were old enough to have
been part of the It community almost from the start. I had been a programmer
in the USA in the 1960s and among other things recall written labelling routines
for such nine-track tapes.
In 1994 Lovdata switched to Unix servers and we thought this would be
an excellent time to migrate from the use of industrial tapes to the more
convenient tape cassettes. No such luck. the tape cassette offered by the
68 | 69
01_2007_5222_txt_ML.indd 69 6-12-2007 15:13:45

Publications Ofice was not supported on our new Alpha machines, and we
had to keep on providing life support for our old system just to make use of its
nine-track tape reader.
As far as I remember it was not until the y2K conversion that we started
receiving data on CD-ROms.
thE yEARS 1985–90
we have all had problems with the production and printing of our legal
gazettes. So apparently had the Publications Ofice. when we received the irst
CELEX material in 1990 we noticed that the full text was missing in many of
the Sector 3 documents from 1985–90. Apparently this was because the Oficial
Journal producer during these years had used a format which could not
easily be loaded into CELEX. however, the full text was added later on.
OVERfLOw DOCUmENtS
In 1990 all the CELEX documents were laid out on tapes, in neatly separated
64K blocks. Of course this was not a big practical problem, we could just
concatenate the blocks, or so we thought.
If a document extended beyond one block, the second and third blocks,
and so on, were called overlow documents. Because we had not read the operations
manual carefully, we presumed, totally without cause or reason as it
turned out, that the overlow documents would tag along with the main document
if this was changed. But of course the Publications Ofice was the epitome
of reason and parsimony and did not send along overlow documents if
they had not changed. thus if a citation 10 digits long (31989S0322) was
wrong in the main document and was replaced by another 10-digit-long citation
(31989S0323), this document was sent along on an update tape, but because
the length of document was maintained, there was no need to send any
overlow documents which were unchanged. however, if the correct citation
was 14 digits long (31989S0322(01)), four new characters were added and all
the subsequent overlow documents would subsequently happily appear in
the update tape.
we realised this when more and more decapitated documents began appearing
in our database, more often than not ending in the middle of a word,
leaving the user in total bewilderment about the intentions of the European
institutions. Even if the original reason for establishing our CELEX database
had been to provide more information on Community law, I will not suggest
01_2007_5222_txt_ML.indd 70 6-12-2007 15:13:45

MEETING OF THE COUNCIL WORKING PARTY
ON LEGAL DATA PROCESSING
that Lovdata is the reason the referendum in 1992 went badly. But surely it was
not one of our more glorious moments.
we had to reorder a completely new stack of tapes and start from the beginning.
the overlow documents also haunted us in another way. Some of them
appeared as ghosts — empty of all content except for a few ields. we theorised
a lot about their meaning — what were they, conceived but not yet
born? — and we waited eagerly for the Publications Ofice to give their bodies
content and meaning. what we discovered was that many of them were indeed
ghosts — either deceased, that is their validity expired before 1 July 1979,
or existing only in another universe, that is they existed only in another language
version. And then some of them did indeed acquire texts after a while.
fILE NAmES
Another thing we wondered about early on was the ile names. they did
not impart much information about what was in the ile, but they were all
equally long. Again I suppose the reason, as for equally long document blocks,
was hidden in the past. And again the y2K conversion brought with it ile
names based on the CELEX document number, and consequently of different
lengths.
CONCLUSION
Of course a lot has changed since the 1990s. first we had the y2K conversion.
And now, just before I left Oslo for this meeting, we received updates
where the documents were coded in XmL and we could fetch them ourselves
with ftP. this is good news and an additional example of the excellent service
and professionalism of the Publications Ofice.
Even if I have permitted myself to highlight some examples of the curious
things that happen in all computer operations, and which we have thus
also experienced as a CELEX host, I cannot emphasise strongly enough the
valuable service which the Publications Ofice has provided over all these
years. this very professional service has not only been invaluable to the EU
member countries, but also to semi-outsiders like Norway. I can only end by
extending a heartfelt thanks to the Publications Ofice for its support and
service all this time.
And lastly, congratulations on your irst 25 years.
70 | 71
01_2007_5222_txt_ML.indd 71 6-12-2007 15:13:45

01_2007_5222_txt_ML.indd 72 6-12-2007 15:13:45

CONCLUSIONS
first, I would like to thank all the speakers who have made this morning
possible. they all showed that we always come back to one central point:
the user. Among the words very often heard this morning, I would mention:
the user, the reuser, and dissemination or redissemination of information and
documents on EU law.
the target user is the citizen, a personality very dificult to deine. And this
is particularly true if we admit that professional users — lawyers and documentalists
in national administrations, in courts, in law irms, in business, in
the academic world, in NGOs and in the EU institutions — are also citizens.
the morning showed that there are some users of expert search in the
room. Expert search has always been felt to be very dificult, and reserved for
some professionals. But thanks to opportunities for getting to know users, the
Publications Ofice has discovered that expert search is not only essential to
particular services within the EU institutions but also to a lot of users outside.
most of the people using EUR-Lex already feel that the simple search tool is
very complete and corresponds to the advanced search in most other systems.
It has also become obvious that the EUR-Lex user needs adequate background
knowledge to search and interpret the list of results, to stream down
the documents and to follow the evolution of EU law.
Discussions have shown that what is needed is a highly professional tool,
even at a time when we strive for transparency and to provide information to
citizens. Giving ordinary people access to a more sophisticated research tool
shows that its use is not restricted to professionals. But there is still a long way
to go to meet the special needs of non-lawyers. It is worth noting that more
and more citizens are becoming more and more interested in law. the number
of users without any legal training using legal information services seems to be
growing. the media increasingly provide legal information: a Luxembourgish
daily newspaper recently had an article about the digitalisation and codiication
of Luxembourgish labour law on its front page.
to make things easier for ordinary people, the EUR-Lex team selects interesting
documents the moment they are put online and constructs iles on
PASCALE BERTELOOT
72 | 73
01_2007_5222_txt_ML.indd 73 6-12-2007 15:13:45

common subjects, mainly because they feature in public discussion, often in
several member States simultaneously or even in all member States. One example
is ‘energy’.
further relection will be required on how best to meet the needs of lay
people and combine them with the needs of professionals, and various solutions
may need to be explored: these might not fall within the sole competency
of the Publications Ofice. Building up ‘corners’ — on consumer protection,
for example, with a list of documents on the subject and a general presentation
— is a way of approaching the lay user which requires a triangular
collaboration between the Publications Ofice and the Directorates-General
for Communication and also, in this case, health and Consumer Protection.
Another way would be to present a highly topical document that was attracting
a lot of comment in the European media. there again, collaboration with
the Communication DG would be necessary.
the responsibility for the management of the system gives opportunities
for contacts with users through the help desk. Contact with users in the member
States also takes place through the Council working Party on Legal Data
Processing. for the EUR-Lex team, national delegates are go-betweens to reach
users in national administrations, but also the national public, and to become
aware of their needs.
It is, and it will always be, the right time to think of evolutions of the legal
information system concerning EU law: stakeholders and the public should
take part, as of course should the member States through their delegates in the
working Party on Legal Data Processing.
01_2007_5222_txt_ML.indd 74 6-12-2007 15:13:45

En tant que déléguée de la Grèce au groupe de travail “Informatique juridique”
et ayant suivi régulièrement les travaux de ce groupe au sein du Conseil
des ministres, je me réjouis d’avoir la possibilité de fêter les 25 ans du droit
communautaire en ligne et plus de 25 ans d’existence de ce groupe et surtout
de son travail fructueux et eficace. WORKSHOP
C’était en 1983 que mon pays, la Grèce, se réjouissait du nouveau système
Celex, prometteur et qui nous avait ébloui. 24 NOVEMBER
J’écrivais alors à l’époque, dans une
revue juridique grecque , combien il était nécessaire de disposer d’un système
donnant accès à toute l’information ju ridique et judicaire, tant pour les services
publics et les universités que pour les particuliers qui s’intéressent au droit
et à son application. Cette nécessité de recherche documentée est devenue
encore plus aiguë en droit communautaire vu le nombre d’actes adoptés, la
diversité des secteurs couverts, la nature spéciale de l’ordre juridique de chaque
état membre, ainsi que la diversité des langues oficielles.
Par sa résolution du 26 novembre 1974 concernant la documentation juridique
automatique, le Conseil des ministres de la Justice avait lancé l’idée de
la création d’un système communautaire de documentation juridique automatique
ouvert à tous les états membres, puisque ceux-ci devaient appliquer le
droit communautaire.
Dans l’article, je me référais aussi au champ couvert par le système, qui
concernait non seulement les textes des traités, mais également le droit dérivé,
les accords conclus par l’Union européenne (à l’époque la Communauté européenne),
la jurisprudence de la Cour de justice ou encore les actes préparatoires.
Ce qui était étonnant, c’était que le système s’étendait aussi aux actes
adoptés par les états membres pour mettre en œuvre le droit communautaire
au sein de leur ordre juridique interne et mentionnait déjà la jurisprudence
nationale. Le système incluait des actes depuis 1952 et la mise à jour se faisait
avec un délai de 3 à 4 semaines à partir de la publication de l’acte. Chaque institution
communautaire était chargée des secteurs le concernant. Le Conseil
fournissait ainsi des données concernant les traités, les relations extérieures, le
droit dérivé, le droit complémentaire. La Commission, le Parlement européen
et le Comité économique et social fournissaient les actes préparatoires. La
Cour fournissait la jurisprudence et le Parlement les questions parlementaires.
La procédure de chargement COURT débutait par OF un formulaire AUDITORS en deux parties.
Une partie, assez détaillée, comprenait plus de 30 rubriques de données documentaires,
et une deuxième présentait le texte entier ou un résumé de celui-ci,
LUXEMBOURG
selon le schéma du formulaire. On pourrait mentionner certains types de ru-
01_2007_5222_txt_ML.indd 75 6-12-2007 15:13:45

01_2007_5222_txt_ML.indd 76 6-12-2007 15:13:45

LEGAL XML — USE OF XML
FOR THE PRODUCTION AND DISTRIBUTION
OF OFFICIAL GAZETTES ( 1 )
mANAGEmENt REPORt
this paper contains the inal report on the work carried out by the working
group ‘XmL: Common use of XmL for the production and distribution of
oficial gazettes’ (‘wG XmL’).
the working group was given the task of proposing a vocabulary in the
form of an XmL schema which simpliies the development and use of XmL
models across the different oficial gazettes in which legislative documents
have to be published.
Analyses by the working group show that a common vocabulary covering
both structure and metadata would not be recommendable. A common vocabulary
on structure would not be successful for two different reasons.
1. Important differences in legislative culture cause different structures in legislative
documents.
2. the advanced status of XmL-based projects in different countries leaves
no possibility to introduce a new schema on structure.
On the other hand, a common vocabulary regarding metadata is proposed
by the working group, while the analyses showed that metadata were highly
congruent across legislative systems, and none of the existing metadata standards
were appropriate for the juridical description of documents.
the working group has created a metadata glossary with agreed deinitions,
and a common XmL schema and document-type deinitions (DtD) were written
based on this.
two obvious beneits are stressed in the report. the common metadata
can be used as an inspiration or a checklist to ensure that the most typical
subjects of legislative metadata are covered, and implementing the common
SØREN BROBERG NIELSEN
Head of Legal
Information Division,
Civil Affairs Agency,
Denmark
( 1 ) the published version is the inal report of the working group on XmL set up by the
European forum of Oficial Gazettes, which was presented between 14 and 16
September 2006 in Ljubljana. the report is co-authored by holger Bagola, Publications
Ofice. the full version of the report, including all its annexes, is available on the forum
website (http://circa.europa.eu/irc/opoce/ojf/info/data/prod/html/index.htm).
76 | 77
01_2007_5222_txt_ML.indd 77 6-12-2007 15:13:47

metadata will facilitate the creation and maintenance of cross-system portals
giving access to multiple legal information systems from the same website.
1. INtRODUCtION
wG XmL was founded in the context of the European forum of Oficial
Gazettes. the working group’s basic mission is described as follows:
‘the XmL project aims that the cooperation between national and
European oficial gazette publishers in the It sector should lead to a
vocabulary which simpliies the development and use of XmL models
which are speciic for legislative documents. this type of vocabulary — in
the form of an XmL schema — would contain models for common objects
but would also allow for deinitions which take account of particularities
in any given national legislation.’
Starting from this point of view, the task of the working group was the
development of XmL models in the form of XmL schema element deinitions.
these models are both related to the description of the document structures in
legislative acts and to the markup of metadata information.
the working group is composed of representatives from these countries:
the Czech Republic, Denmark, Germany, Greece, Estonia, the former yugoslav
Republic of macedonia, france, hungary, Italy, Lithuania, the Netherlands,
Portugal, Romania and Sweden, as well as the European Union. Other countries
having expressed an interest in being informed about the results of the
working group are: Austria, Belgium, Ireland, finland and the United Kingdom.
In the beginning the working group was chaired by Nina Koch (Denmark).
however, because of other tasks in the context of the forum, she passed
the chair to Søren Broberg Nielsen (Denmark).
In order to get to know the members of the working group, a questionnaire
asking for their experiences and expectations was sent out to the designated members.
the answers show that at least basic experiences are present concerning
XmL-based methodologies and technologies. But regarding the two techniques
for designing XmL-based grammars — DtD and XmL schema — it becomes
obvious that most experiences are still limited to the DtD variant. this impression
gets clearer when expectations for the following are taken into account:
• the design of legal documents, which concentrates on the granularity used
for the markup of the different document components,
• the exchange of experiences, which expresses the wish to discuss different
approaches and to learn from the successes and problems met in different
steps of the work,
01_2007_5222_txt_ML.indd 78 6-12-2007 15:13:47

WORKSHOP
• publishing technologies which are based on, or proit from, the use of
XmL-based markup, and
• technical and organisational work with XmL, which includes experience
in the use and coniguration of tools.
four meetings were organised (see Section 2.1.). the minutes of them are
available in the ‘members area’ (‘Ongoing projects’ section) of the forum website
(http://circa.europa.eu/irc/opoce/ojf/info/data/prod/html/index.htm).
the report is structured in different parts. After some information on the
use of XmL in the production of oficial gazettes in the various countries, a
description of the work undertaken gives detailed information on the meetings,
the subjects discussed and their interests. Contacts with other working
groups are summarised afterwards.
Chapter 3 gives a general overview of the discussions which led to limitations
of the initial mission and the deinition of the work to be executed by the
working group. Chapters 4 and 5 describe the results of the work, a glossary of
common metadata and details of its technical implementation as a w3C XmL
schema and a DtD. the last chapter underlines the beneits of the use of a set
of common metadata.
the report is completed by the metadata vocabulary, the schema, the
DtD, and instructions on how to integrate the grammars and where to access
them. the last annex contains a list of all members taking part in the working
group meetings.
In the very irst meeting, representatives of the countries, when talking
about their different experiences in the work with XmL, made it clear that they
all could proit from the exchange of experiences. the lack of intuitive and user
friendly interfaces to the XmL authoring tools was identiied as one of the
most serious problems for XmL projects.
Countries
EU
Italy
XML
from 1 may 2004, the Official Journal of the European Union
will be composed using XmL markup (‘formex v.4’,
formex being ‘formalised Exchange of Electronic Publications’;
the ofice previously used SGmL (since 1985:
formex V.1, V.2, V.3).
the NORmA editor and the NiR editor were developed in
the context of the Norme-in-rete project. the two editors
thus became the main editing tools in the Italian legislative
process. the NORmA editor is an add-on to microsoft
word, and the NiR editor is based on Java and XmL technology.
has good expertise on use of XmL.
78 | 79
01_2007_5222_txt_ML.indd 79 6-12-2007 15:13:47

Sweden
Denmark
Germany
france
former
yugoslav Republic
of macedonia
Netherlands
Romania
Earlier adopter of XmL. Its irst project failed due to the
reluctance of the users to deal with XmL tags. Learning
from their mistakes, the Swedish team presented a new
prototype which is proven to be much more user friendly
than the irst one and acceptance by the users could be foreseen
soon.
Denmark has considered using the Italian NiR editor in the
LexDania project. But in the irst phase of the project a
server-based transformation from .doc-iles to Lex Dania
XmL will be used.
Bundesanzeiger print: partly XmL based; eBundesanzeiger:
XmL based; Bundesgesetzblatt: partly XmL based; use of
DtD.
Use of XmL and schema. Relies on dematerialisation or
digitised documents in its production process.
A production system is under development. It is completely
based on XmL and also allows for the charging of information
in an information system on legislation.
A good experience of XmL. Develops an XmL-based tool to
manage the consolidation process. Special interest was
drawn to the consolidation facilities in that system, which
allows extracting a document in the particular form at any
date in its life.
A production system based on XmL is under development.
It is well known that other countries also have important developments of
production and information systems in this context. therefore it was rather
regrettable that no reports on experiences were available.
2. wORK UNDERtAKEN
2.1. Meetings
four meetings were organised in 2005 and 2006. they were all hosted by
the Ofice for Oficial Publications of the European Union in Luxembourg.
the irst meeting took place in march 2005, and considered the following
main items:
• presentation of members,
• the XmL view on legal documents,
• presentation of the Circa discussion forum,
• structure of European and national legislative acts,
• transformation of European directives into national law,
• terminology and metadata.
the second meeting, in June 2005, concentrated on the following items:
• status of XmL in the different countries,
• report on the Swedish project,
01_2007_5222_txt_ML.indd 80 6-12-2007 15:13:48

WORKSHOP
• NiR and the NiR editor,
• discussion of editorial tools,
• presentation of the Dutch legal information system and the consolidation
approach,
• metadata.
the third meeting, in march 2006, took into account the following items:
• report on the Copenhagen meeting of the forum,
• presentation and discussion of the working plan for 2006,
• continuation of work on the metadata vocabulary,
• discussion of a namespace or inclusion-based approach for the integration
of the metadata speciications in local applications,
• presentation of the Eurovoc thesaurus,
• reports on metadata indexation in the various countries.
the fourth meeting, in June 2006, was supposed to be the last one, and
had the following items on the agenda:
• report on news from the forum,
• reports on news from the various countries,
• presentation of the N-Lex system, which could possibly proit from the
deinition of common metadata,
• presentation of the current status of the metadata vocabulary,
• discussion of outstanding issues on the metadata vocabulary, deinition of
cardinality of elements, naming conventions, versioning,
• presentation and discussion of a draft table of contents for the inal report.
In the last meeting it was also proposed to develop a prototype system
which should illustrate the beneits of the use of a common metadata vocabulary.
2.2. Subjects
As previously stated, one of the common problems concerned the availability
of tools which offer good interfaces between the human operators and
the XmL instances. As most of the human operators are not experts in XmL
issues, but specialists in drafting legislation, it is obvious that they should not
be confronted with XmL markup. the different presentations, as well as the
experiences of users, showed that no system can cover all needs. In most cases
editing tools will have to be conigured for local needs, or speciic applications
will have to be developed.
80 | 81
01_2007_5222_txt_ML.indd 81 6-12-2007 15:13:48

the development of the metadata vocabulary was based on a synopsis of
the metadata which are used in the different systems in the different countries.
It turned out that the items could be classiied in two groups. the irst was
related to the national or local legislation system. these metadata can hardly
be reused on a more generic level. the other group, however, contain elements
which are present in more or less all of the mentioned systems.
the discussion also revealed that the development of a common metadata
vocabulary has to proceed on two levels. A glossary has to be deined which
does not only list the terms to be used, but also gives deinitions of what is
meant by a speciic item. Only then can the terms be used to create an XmL
grammar for the metadata concerned.
A special item of the discussion concerned the description of the contents of
legal acts within the metadata section. the use of thesauri and/or ontologies has to
be taken into account. the multilingual approach such as presented by Eurovoc is
beyond the scope of most of the national systems. But it was underlined that, for
the domains covered by Eurovoc, this approach is useful. It has, of course, to be
completed by additional information on those subjects which are not covered by
European legislation and as a consequence not taken into account by Eurovoc.
Another subject concerns the reuse of metadata information by systems
giving access to legislative documents. the N-Lex system, which is still experimental,
could certainly be improved and simpliied if documents foresee a
commonly readable description of their contents.
2.3. Interests
In the meetings of the working group, some issues more or less related to its
mission have emerged. members of the working group showed interest in many
different aspects of the use of technology in all parts of the production of oficial
gazettes. the interests expressed by the members was often rooted in challenges
experienced in the different countries and, as often in a working group, others
have experienced the same challenges. the group thus showed its usefulness by
giving the opportunity to exchange common challenges and solutions.
Among the many and diverse interests expressed by the members, two
issues proved to be of general interest, namely:
• editorial tools for drafting legislation,
• automatic indexation of legislation.
the working group chose different approaches to the issues. Some members
had an extensive knowledge of developing and/or implementation of edi-
01_2007_5222_txt_ML.indd 82 6-12-2007 15:13:48

WORKSHOP
tors, thus providing the opportunity to examine the issue in one of the meetings.
however, members did not have suficient expertise regarding automatic indexation,
leading them to recommend the founding of a new working group.
As mentioned, all the projects in the working group had considered the
issue of an appropriate and user-friendly editor for drafting legislation in an
XmL environment. One of the main grounds for concern about editors is the
more and more widespread wish for capturing the legislative text at the source.
If drafters write the legislative text in XmL, the text is ready for integration into
information systems, publishing and archiving procedures, and there is no
need for additional markup in prepress production — additional markup being
both costly and creating opportunities for mistakes and errors.
the working group examined the case of editors in its second meeting,
with a presentation and demonstration of the NIREditor which was developed
in the context of the Norme-inrete project.
the working group’s discussions showed that there are at least three solutions
to the editor question. these are:
• a speciic editor (such as the NIREditor),
• a generic XmL editor such as Xmetal which was used in the Swedish
project and which can be conigured in such a way that all tags are hidden
for the user, and
• a generic word processor (e.g. microsoft word) with additional processing
into XmL, such as the EU’s Legiswrite.
the discussion showed that there were various arguments for and against
the different approaches, and therefore no speciic conclusions could be drawn.
here are some of the arguments:
‘A specific editor will only allow valid XML with no use for an a posteriori
validation to the annoyance of the drafter’
‘It is easier to do additional processing with the legislative texts than
introducing a new editor to the users’
‘The cost to develop and support a specific editor is too high’
‘Cost of software licence for a generic XML editor is too high, if the
editor should be used by all drafters’
‘A specific editor gives the possibilities to aid the drafter with legal
techniques at a much higher level’
82 | 83
01_2007_5222_txt_ML.indd 83 6-12-2007 15:13:48

In the third meeting, the working group discussed indexation vis-à-vis
common metadata representing indexation. to introduce the aspects of indexation,
a presentation of the Eurovoc system as a multilingual thesaurus for
the indexation of legislation was given by the Publications Ofice. During the
discussion the usefulness was clearly underlined, but it was also stated that
Eurovoc in its current version does not cover all domains of national legislation,
but concentrates on European needs. many working group members
were surprised at how time-consuming a proper and thorough indexation is,
both regarding the time spent on every document and the overheads covering
training of staff and administration of the thesaurus.
the discussion inspired the delegation from the Czech Republic to propose
a comprehensive examination of the issue of automatic indexation of
legislation within the working group. It was, however, decided that the issue
was beyond the scope of the XmL working group, but so interesting and potentially
useful that the chair of the European forum of Oficial Gazettes proposed
the foundation of a new working group concerned with automatic indexation
of legislation.
2.4. Contact with other working groups
two of the forum’s working groups cover issues that are of interest to the
XmL working group, and it was decided to contact them concerning
authentiication and data capturing. the aim of these contacts was to give
knowledge to the other working groups so that they could take the XmL dimension
into account during their work.
Regarding the working group on ‘Data capturing at the source’, two questions
were asked.
• Is it feasible to introduce XmL in data capturing at the source?
• If so, what are the means?
Regarding the working group on ‘Electronic publication of legislation:
methods of authentiication of the texts’, attention was drawn to two aspects
of XmL and authentiication.
• If XmL is chosen as the master document in the production of the electronic
legal gazette, what is the legal status of the master document? Is the
XmL master document the authentic document or is it the XmL document
combined with the presentation?
• what are the possibilities for authentiication in the XmL world?
01_2007_5222_txt_ML.indd 84 6-12-2007 15:13:48

WORKSHOP
3. SCOPE
the expected result from the initial mission of the working group was an
XmL vocabulary for the description of the structure of legal acts in the different
countries. One of the ideas behind this objective was the idea to avoid the
development of different XmL grammars in the various countries.
the starting point of the discussion was a comparison of the markup in
three different acts: from the European Union, from Italy and from Denmark.
A irst glance at legal acts in different countries proves that there are obvious
similarities, but also important differences. the comparison of the document
structure at the highest level makes this situation evident. the following synopsis
shows the markup chosen in Danish, Italian and European laws.
DK IT EU
Dokument LinguaDoc
meta+ titelGruppe
DokumentIndhold
UnderskriftGruppe
Bilag+
DecretoLegislativo
meta intestazione
formulainiziale
articolato formulainale
conclusione annessi?
ACt BIB.INStANCE
tItLE PREAmBLE
ENACtING.tERmS
fINAL
the differences mainly mean that corresponding elements are absent in
other bodies of legislation. the number of these differences is even bigger when
details are taken into account. the following example illustrates this situation:
It articolo = EU article
≠ DA artikel
On the other hand it is also clear that up to a certain extent, but only theoretically,
markup systems can be exchanged between countries if the chosen
names have a meaning in the other country.
Examining the different approaches, it becomes obvious that the markup
on the contents level is done on two levels — a semantic markup for the main
structural document components and a more generic markup of elements on a
deeper level. A common approach, however, could still be realised by using
the following approach.
first of all, for the naming of elements, semantic names have to be dropped
and replaced by a system of recursive containers. Semantic information could be
maintained on the attribute level. the advantage is that all validation rules are
independent of a given language, so the control mechanisms could be reused in
all countries. National speciic names could be integrated by means of predeined
grammar fragments, which are included when the DtD or schema is used.
84 | 85
01_2007_5222_txt_ML.indd 85 6-12-2007 15:13:48

Although this method could be a solution, it will certainly not meet all the
needs which exist in different countries. however, the problem of indicating
semantic correspondences is not solved; it is just transferred to another level.
As the values for the attributes have to be deined outside the core grammar,
the replacement of the DtD by schemas is nearly mandatory, especially because
the conventions for creating a system of attribute values are rather limited
for a DtD. Another fact concerns the readability for a human user. In
general, it is said that XmL instances are made for machines, and the markup
as shown in the following example only bears out this opinion.
VERORDNUNG (EG) Nr. 362/2005 DER KOMMISSION
vom 3. März 2005
zur Ablehnung von Anträgen auf Erteilung von
Ausfuhrlizenzenim Getreidesektor für Erzeugnisse des
KN-Codes 1101 00 15
DIE KOMMISSION DER EUROPÄISCHEN GEMEINSCHAFTEN —
gestützt auf den Vertrag zur Gründung der
EuropäischenGemeinschaft,
gestützt auf die Verordnung (EG) Nr. 1784/2003 des
Rates vom
29. September 2003 über die gemeinsame
Marktorganisation fürGetreide (1),
01_2007_5222_txt_ML.indd 86 6-12-2007 15:13:48

WORKSHOP
gestützt auf die Verordnung (EG) Nr. 1342/2003 der
Kommissionvom 27. Juli 2003 mit besonderen
Durchführungsbestimmungen überEinfuhr- und
Ausfuhrlizenzen für Getreide und Reis (2),insbesondere
auf Artikel 8 Absatz 1,
in Erwägung nachstehenden Grundes:
Die Anzahl der Anträge auf im Voraus
festgesetzteErstattungen für Erzeugnisse des KN-Codes
1101 00 15 istbedeutend und von spekulativem
Charakter. Es sollten deshalballe Anträge abgelehnt
werden, die am 1. März 2005eingereicht wurden —
HAT FOLGENDE VERORDNUNG ERLASSEN:
Artikel 1
Gemäß Artikel 8 Absatz 1 der Verordnung (EG) Nr.
1342/2003wird die am 1. März 2005 beantragte Erteilung
von Lizenzenfür die Ausfuhr von Erzeugnissen des KN-
Codes 1101 00 15abgelehnt.
Artikel 2
Diese Verordnung tritt am 4. März 2005 in Kraft.
86 | 87
01_2007_5222_txt_ML.indd 87 6-12-2007 15:13:48

Diese Verordnung ist in allen ihren Teilen verbindlich
und giltunmittelbar in jedem Mitgliedstaat.
Brüssel, den 3. März 2005
Für die Kommission
J. M. SILVA RODRÍGUEZ
Generaldirektor für Landwirtschaft und Entwicklung
desländlichen Raumes
furthermore, the information on XmL-based projects in the different
countries shows that XmL is used in nearly all of them. the development of
grammars is inished or at a very advanced level. So for the development of a
common vocabulary, problems might arise from the fact that resources have
already been allocated, and replacing the home-made solution which corresponds
exactly to the needs will hardly be welcome.
the development of a vocabulary for common structural markup does not
seem, therefore, to be feasible because of important differences in legislative
culture on the one hand, and the advanced status of XmL-based projects in the
different countries on the other hand. the discussions, however, also showed
that there is an important conformity on the metadata level. A synoptic inventory
listed the various metadata elements which are in use in the various countries.
On this basis a list of about 30 elements could be identiied, which are of
interest in the various legal information systems.
the list also made evident some terminological dificulties. therefore it
was decided to create a glossary with deinitions for each term which is part of
the metadata element collection. the glossary was completed by some terms
01_2007_5222_txt_ML.indd 88 6-12-2007 15:13:48

WORKSHOP
which describe general components of an act, but do not belong to the metadata
elements.
On the XmL level, the grammar was built on the basis of w3C XmL
Schema technology. for each of the identiied elements an XmL model was
deined. A supplementary container was created to keep the metadata information
together and to add cardinality to the elements. In the end, most of the
elements turned out to be optional, either because the information is not available
in all documents or because it is not used in some countries.
As already mentioned in the introduction, some countries have not yet
taken the decision to migrate their XmL-based systems to schemas; they maintain
a DtD-oriented approach. this is why the schema is accompanied by a
corresponding DtD. It has to be kept in mind that the DtD was derived from
the schema and cannot cover all the features offered by a schema grammar.
for the implementation of the schema three methods were presented:
• copying the contents of the metadata schema into the local schema,
• including the metadata schema by means of the xs:include element,
• deining a namespace and calling the metadata schema by means of the
namespace location address (xs:import element).
the irst approach requires a lot of management in case the metadata schema
evolves. So it was no longer followed. whether the metadata schema
should be included or imported depends on the organisation of the local schema.
therefore it was decided to maintain both solutions. As it is not possible
to refer to the same schema for inclusion and import, a container schema was
created which in fact includes the original schema. the namespace reference in
the local schema has to be linked to this container. for the use of the schema
within these instances this method remains transparent.
According to the XmL standard, the decision to use a DtD is a decision
against namespaces, as they are not supported within a DtD. So the metadata
DtD can only be included by means of a public or system reference call.
the import or inclusion of a schema or a DtD is only executed when the
grammar is opened by a coherent tool, for example a parser. A conditio sine qua
non for this approach is the online availability of the metadata schema and
DtD. therefore both grammars will be accessible at the same address as the
formex schema which is used for the documents published in the Official Journal
of the European Union.
In order to demonstrate how useful the common metadata vocabulary can
be, a prototype will be developed. Documents from different countries on the
same contents domain will be searched and presented to the user. he or she
88 | 89
01_2007_5222_txt_ML.indd 89 6-12-2007 15:13:48

will then not have to know the local methodologies applied for the integration
of metadata information.
4. mEtADAtA VOCABULARy
the need for metadata for the description of document contents is in fact
not new. International standards for the exchange of bibliographic information
and/or library catalogues have been created and are used regularly.
One of these standards, perhaps the most important one, is mARC21,
machine-Readable Cataloguing, which was developed and is maintained by
the American Library of Congress. It offers a records-based system with rather
a large variety of metadata ields. the XmL adaptation of this standard is not
very satisfying because it resembles the abovementioned approach for a common
markup, as can be seen in the following illustration.
01142cam 2200301 a 4500
92005291
DLC 19930521155141.9 920219s1993 caua j 000 0eng
92005291 0152038655: $15.95 DLC DLC DLC lcac
…
A poem about numbers and
theircharacteristics. Features anamorphic, or
distorted, drawingswhich can be restored to normal by
viewing from a particularangle or by viewing the
image’s reflection in the providedMylar cone.
Arithmetic Juvenile poetry.
Children’s poetry, American.
01_2007_5222_txt_ML.indd 90 6-12-2007 15:13:49

WORKSHOP
Arithmetic
Poetry.
American poetry.
Visual perception.
Rand, Ted, ill.
Another standard is derived from mARC. mEtS (metadata Encoding and
transmission Standard) was also developed by the Library of Congress. It consists
of 34 ields and its objective is as follows.
‘the mEtS schema is a standard for encoding descriptive, administrative,
and structural metadata regarding objects within a digital library expressed
using the XmL schema language of the word wide web Consortium. the
standard is maintained in the Network Development and mARC Standards
Ofice of the Library of Congress, and is being developed as an initiative
of the Digital Library federation.’
An interesting fact is that it is based on an XmL grammar.
Another standard is frequently used in the context of XmL-based document
markup, although it was not originally based on or limited to XmL.
DCmI (Dublin Core metadata Initiative) offers 15 ields for the collection of
metadata. they are often used as markup within the resource description
framework (RDf) container system.
Rose Bush
A Guide to Growing Roses
Describes process for planting and
nurturing
different kinds of rose bushes. 2001-01-20
It is obvious for this approach that 15 ields might not be suficient for
certain projects. this is why DCmI offers the possibility of extending the ields
90 | 91
01_2007_5222_txt_ML.indd 91 6-12-2007 15:13:49

y the deinition of subields. One project which proceeded in this way is
mIReG (managing resources for e-government). mIReG is the result of an IDA
project between the United Kingdom and the EU.
Not all these approaches are appropriate for the juridical description of
documents. At the irst meeting of the working group a list of metadata ields
was identiied and subsequently completed (see the synopsis in Annex 1). this
list has the following features.
• Some of the mentioned ields are used in various countries under more or
less the same name.
• Other ields are speciic to a certain country, and the concept seems not to
be taken into account elsewhere.
• Some of the deinitions overlap and need to be clariied.
In order to analyse how far existing standards could be used for the metadata
ields collected in the synopsis, DCmI was applied, leading to the following
results:
Metadata fields collected in the synopsis
DCMI
Identiier on metadata level
identiier
Information (issue number, issue name, approval date, printing date) identiier
Reference to oficial journal
relation
Subject
subject
Act ./.
Additionally used abbreviation ./.
Authors service (ministry)
creator
Code number ./.
Consolidated or not
type
Contents
description
Date of adoption/signature
date
Date of application
date
Date of decision
date
Date of enactment
date
Date of entering into force
date
Date of expiration
date
Date of publishing/publication
date
Date of repeal
date
Date of the law
date
Document type ./.
Keywords (subject of the document)
subject
Language of the document
language
Nickname of the legal document (subject, friendly name) ./.
Number used in the ministry archives ./.
01_2007_5222_txt_ML.indd 92 6-12-2007 15:13:49

WORKSHOP
Metadata fields collected in the synopsis
DCMI
Oficial abbreviation ./.
Original vs amendment ./.
Preamble ./.
Reference to basic legislation
relation
Reference to changing text (if consolidated)
relation
Reference to directive (if applicable)
relation
Reference to draft
source
Reference to modiied acts
relation
Reference to next version (if consolidated)
relation
Reference to previous version (if consolidated)
relation
Reference to source (if consolidated)
relation
Reference to translation
relation
References
relation
Region code ./.
Reprint or not
type
Signature
creator
Status (in force or not)
type
Summary ./.
title
title
Some of the ields are indeed covered by corresponding descriptors in
DCmI. however, most of these descriptors have to be used more than once,
so the real meaning of a metadata ield is no longer evident. Other ields are
not covered at all. As a consequence the need for a speciic markup system for
metadata in legal documents became obvious.
So it was decided to identify relevant metadata ields and to deine their
role within the juridical analysis of the object.
5. tEChNICAL SOLUtION
the metadata glossary has been transformed into an XmL schema (see
Annex 2A) and a parallel DtD (see Annex 2C). Both grammars are mutually
interchangeable. the parsing with the schema, however, gives more details on
the correctness of the instance. this becomes evident when comparing the
models for the CELEX number: in the DtD it is only a sequence of characters,
while in the schema a pattern is deined which helps to control the syntax of
the codes while parsing.
the names of the deined elements all start with the preix ‘efog.’. this
guarantees that the elements may always be identiied within the local
92 | 93
01_2007_5222_txt_ML.indd 93 6-12-2007 15:13:49

grammar or in the instances. the elements are deined in alphabetical order,
just to simplify the access to the elements for a human user.
the various models are always completed by annotations or comments
which explain the semantic role of a given element. these explanations are
based on the deinitions in the metadata glossary.
Although it is possible to use the various elements at any place within an
instance, it is recommended to integrate them within a single block. this
should simplify searches in the documents. furthermore a control of cardinality
is only possible within the predeined block. Another possibility is the creation
of an element in the local schema which is related to the efog-type. from
a technical point of view this approach is as eficient as the use of the efogblock
element; the only disadvantage is that for researching the instances the
name of the block should be known.
the grammars will be available from the formex website, which is managed
by the Publications Ofice (http://formex.publications.europa.eu/schema/).
the URL is completed by the name of the grammar and is also valid for
the DtD. the directory will always contain all versions of the grammars. So in
the case of improvements there is not necessarily the need to review and update
all instances. the syntax of the names of the grammars has been deined
as follows:
• [ 1] name :: = name.schema | name.DtD
• [ 2] name.schema :: = ‘efog_’ ‘cnt_’? date ‘.xsd’
• [ 3] name.DtD :: = ‘efog_’ date ‘.dtd’
• [ 4] date :: = d{8}
• [ 5] d :: = ‘0’|‘1’|‘2’|‘3’|‘4’|‘5’|‘6’|‘7|‘8’|‘9’.
the date relects the date when the version of the grammar was adopted.
the constant part ‘cnt_’ is only to be used when the schema is referred to by
means of the namespace technology. (the schema is to be found in Annex 2B.)
As discussed earlier in this document, the implementation can be achieved
in different ways: a copy-paste method, the inclusion of the schema, or a
namespace reference. for the DtD, the copy-paste method and the inclusion
are feasible. the copy-paste method is highly deprecated; details on inclusion
and namespace references can be found in Annex 2D.
6. BENEfItS Of COmmON mEtADAtA
In this chapter the working group will exemplify the beneits of common
metadata by different cases of use that hopefully will promote the dissemina-
01_2007_5222_txt_ML.indd 94 6-12-2007 15:13:49

WORKSHOP
tion of the common metadata. It is important to understand that the implementation
of the common metadata does not exclude the use of systemspeciic
metadata; with the proposed naming convention, there should not be
any conlicts.
the main beneits of the common metadata could be:
• improved access to a set of standard deinitions of metadata terms through
the Internet,
• improved standardisation of metadata for dissemination and international
comparisons,
• support to XmL structures for searching and exchanging metadata on legislation.
6.1. Checklist for metadata design
when creating or redesigning a legal information system, the common
metadata can be used as an inspiration or a checklist to ensure that the most
typical subjects of legislative metadata are covered. this advantage also covers
the proposed grammar, where metadata designers can use the XmL schema
deinitions when deining local metadata.
6.2. Facilitating access to legislation across different legal
information systems
Implementing the common metadata will facilitate the creation and maintenance
of cross-system portals giving access from the same website to multiple
legal information systems. this is made possible when the portal can be
designed with a uniform interface to metadata in the different systems.
the opposite case is also facilitated when new legal information systems
are created or existing systems are renewed. the new or redesigned sites can
more easily be integrated in the portal; if the portal supports the common
metadata, the access to a metadata search is more or less prefabricated.
Besides the more technical beneits for cross-system portals, users will
hopefully ind it easier and more intuitive to query the different systems. when
metadata are common, the query in different systems will not be hindered by
confusion over ‘where to search for what’.
94 | 95
01_2007_5222_txt_ML.indd 95 6-12-2007 15:13:49

01_2007_5222_txt_ML.indd 96 6-12-2007 15:13:49

ELECTRONIC PUBLISHING OF LEGISLATION —
TOWARDS AUTHENTICITY ( 1 )
1. INtRODUCtION
the working Group on Authenticity was one of the working groups established
by the European forum of Oficial Gazettes in September 2004 in
Vienna.
Since 2004, there has been a clear need for work on this issue. the use of
electronic legal gazettes has been growing rapidly. At the same, the number of
paper copies of legal gazettes and the number of subscribers have been decreasing.
therefore printing the paper copies has become less cost-eficient.
the working group has analysed the general context of authenticity in the
electronic publishing of legislation. the authenticity of electronic legislative
documents cannot be seen as an isolated issue, but rather as a part of the complicated
process of publishing legislation in paper and electronic form.
Since the forum meeting in Copenhagen in September 2005, there has
been ongoing discussion on the authenticity of electronic legal gazettes in
most countries. In the majority of them the paper version of law is still the only
authentic one. the approach of the working group has been very pragmatic,
based on the experiences of the participating countries, which have taken very
different approaches to solve the issue of authenticity.
the methods of authentication of the texts in legal electronic gazettes have
been discussed in the working group by delegates from Austria, Belgium, Estonia,
finland, france, Germany, Greece, hungary, Iceland, Italy, Latvia,
Lithuania, Portugal and Spain. the group has been chaired by Aki hietanen
(finland) with maria manuela Cruz (Ofice for Oficial Publications of the
European Communities) acting as secretary.
AKI HIETANEN
Director of Information
Services, Ministry of Justice,
Finland
( 1 ) the published version is the inal report of the working Group on Authenticity set up
by the European forum of Oficial Gazettes which was presented on 14 and 15 June
2007 in its helsinki–tallinn meeting. During the workshop, mr hietanen presented the
state of work in November 2006. the full version of the report, including all its
annexes, is available on the forum website (http://circa.europa.eu/irc/opoce/ojf/info/
data/prod/html/index.htm).
96 | 97
01_2007_5222_txt_ML.indd 97 6-12-2007 15:13:51

Since the working group was set up, a number of signiicant changes have
taken place.
• Electronic signatures have been introduced in electronic legal gazettes.
• Secure servers have been established for the delivery of electronic gazettes.
• Certiicates have been used with secure servers.
• Legislative reforms have been implemented concerning the status of electronic
legal gazettes.
• Authentic and consolidated law has been made available in electronic
form.
• the number of paper copies of legal gazettes has been cut down.
historically, there are national milestones in the development of electronic
legal gazettes.
• Estonia: 23 January 2001: the paper and electronic versions are equal;
from 1 January 2007: the electronic version became the only authentic
one.
• Belgium: 1 January 2003: the electronic version is not authentic, but is the
only version widely available (ive paper copies are authentic); in addition
a help desk with free-of-charge access is described in the law and a summary
of the legal gazette is available at each court.
• Austria: 1 January 2004: only an electronic version, which is authentic, is
available. In addition, non-authentic electronic versions (htmL/PDf) and
ive paper copies for archiving.
• France: 1 July 2004: the paper and electronic versions are equally authentic.
• Hungary: 1 January 2006: the electronic version of the legal gazette has
been made authentic.
• Portugal: 1 July 2006: the electronic version became the authentic oficial
journal. the acts only become oficial after their publication in the electronic
journal at www.dre.pt. the paper version of the oficial journal has
not been published since 31 December 2006.
• Slovenia: 1 January 2007: the paper and electronic versions are equally
authentic; in case of conlict, the electronic version prevails.
• Denmark: 1 January 2008: the electronic version will become the only
authentic one.
the national reports, meeting reports and other documents of the working
group are available at the forum website.
01_2007_5222_txt_ML.indd 98 6-12-2007 15:13:51

WORKSHOP
the working group has focused on the following questions.
1. what is the relation of the electronic version to the paper version of the
legal gazette?
2. what kind of technical tools are available for the authentication of electronic
legal gazettes?
3. what kind of reliable processes are necessary in the production and distribution
of electronic legal gazettes?
4. what is the quality control in the production chain?
5. which tools and methods of authentication are generic and also could be
used in other countries?
6. what level of reliability and authenticity is suficient?
7. what kind of guidelines and good practices can be spread within the
European forum of Oficial Gazettes?
the authenticity of electronic legal documents has to be analysed in the
framework of information security and general security attributes of electronic
documents. the relevant security attributes of electronic legal documents
are:
• integrity,
• authenticity,
• availability,
• utility (usefulness),
• control.
these attributes of information are non-overlapping, and refer to unique
aspects of information. Any information security breach can be described as
affecting one or more of these fundamental attributes of information.
Integrity refers to being correct or consistent with the intended state of
information. Any unauthorised modiication of data, whether deliberate or accidental,
is a breach of data integrity. for example, data stored on a server are
expected to be stable — they are not supposed to be changed at random by
problems with the operating system. Similarly, application programs are supposed
to record information correctly. One additional attribute of integrity
could be traceability — the possibility to trace any amendments made to the
original texts, in order to guarantee the origin and integrity of the text.
Availability means having timely access to information. for example, disk
crashes and denial-of-service attacks both cause a breach of availability. Any
delay that exceeds the expected service levels for a system can be described as
a breach of availability.
98 | 99
01_2007_5222_txt_ML.indd 99 6-12-2007 15:13:51

Utility equals usefulness. for example, if someone encrypted data on
server or disk to prevent unauthorised access or undetected modiications, and
then lost the decryption key, that would be a breach of utility. the data would
be integral, authentic, and available — they just would not be useful in that
form. the same goes for the storage of data in a format inappropriate for a
speciic computer architecture, for example SGmL instead of XmL, EBCDIC
instead of ASCII or a DVD-ROm instead of a 1.44 mb diskette.
Control of information usually refers to the ownership of data and to the
different possibilities to modify or delete data. the access control to databases
with the ‘original versions’ of legal gazettes is a typical challenge to this aspect
of security.
On the question of authenticity, a distinction has to be made between (a)
the authenticity of electronic documents, (b) the authentication of production
processes of electronic documents, and (c) authentication of printing and delivery
of electronic documents
As regards authenticity and authentication, some principles are essential.
1. the authentication processes should be effective, eficient, reliable and
easy to use.
2. the choice of technology, services and technical solutions should emphasise
compliance with standards.
3. Proportionality should be stressed: the scope and intensity of the practical
measures in the authentication should be in proportion to the degree of
beneits that the institutions and citizens are expected to have.
Often authenticity equals reliability. A reliable document or an authentic
document is one endowed with trustworthiness. Speciically, trustworthiness
is conferred to a document by its degree of completeness and the degree of
control over its creation procedure and/or its author’s reliability. the author’s
reliability means in practice, for example, that a legislative act is suficiently
authenticated if it is signed by the president or the prime minister.
the ISO standard on record management (ISO 15489) deines authenticity,
reliability, integrity and useability in the following terms.
• An authentic record is one that can be proven to be what it purports to be;
to have been created or sent by the person purported to have created or
sent it; and to have been created or sent at the time purported.
• A reliable record is one whose contents can be trusted as a full and accurate
representation of the transactions, activities or facts to which they attest
and can be depended upon in the course of subsequent transactions
or activities.
01_2007_5222_txt_ML.indd 100 6-12-2007 15:13:51

WORKSHOP
• the integrity of a record refers to its being complete and unaltered.
• A useable record is one that can be located, retrieved, presented and interpreted.
In the discussion concerning the quality requirements or criteria for electronic
legal gazettes, the Ofice for Oficial Publications of the European Communities
has analysed,in addition to availability and authenticity, the following
criteria — accessibility, certiication, durability and traceability.
the most suitable means for choosing the type of authentication could be
a cost–beneit analysis, risk analysis or SwOt analysis deining the strengths,
weaknesses, opportunities and threats of different methods. this analysis
would specify the needs of different target functions. Each need would be
evaluated with the four main security criteria: integrity, availability, authenticity
and usability. the target of the analysis would be to identify what you
may win and what you may lose (conidence, trusted process, etc.) and what
is the impact of the choices.
As an example, here is the SwOt analysis.
SWOT ANALYSIS OF AUTHENTICATION METHODS
USE OF ELECTRONIC SIGNATURES
Strengths
— Eicient and reliable methods
for authentication
— Several techniques and standards
available, also open source
signatures
Opportunities
— Electronic signature is applicable
to all legislative documents
— Essential part of electronic
commerce
Weaknesses
— Not necessary if worklow
and secure servers are used
— Diiculty to choose the most
suitable electronic signature
— Diiculty in transferring
the signature to new document
formats
Threats
— The archiving of documents
with electronic signature
is problematic
— The electronic signature has to be
renewed (re-signed) frequently
100 | 101
01_2007_5222_txt_ML.indd 101 6-12-2007 15:13:51

In a similar manner, a SwOt analysis can be made of the use of secure
servers and protocols.
SWOT ANALYSIS OF AUTHENTICATION METHODS
USE OF SECURE SERVERS AND CERTIFICATES
Strengths
— Eicient method for ensuring
the data transfer
— Several techniques and standards
available, also open source
— Data encryption is used
Opportunities
— Experience found in all countries,
well-known techniques
— Secure server easy to establish
and maintain
Weaknesses
— Does nor guarantee the
authenticity of documents
— Diiculty to choose the most
suitable combination of SSL
and protocols
— Progress in data encryption
outdates the old one (40-bit
encryption etc.) —> Risk of
intrusion and hacking
Threats
— The standards and protocols
develop constantly
the countries planning to develop electronic version of the legal gazette
have several possibilities and options. here are some examples.
No specific authentication of the legal gazette
If the producer of paper and electronic version is the same authority, the
electronic version may be deemed to be authentic on the basis of the competence
of the authority.
Use of electronic signatures with legal gazettes
If there is no worklow or if there are different processes for paper and
electronic versions, it may be useful to attach electronic signatures to the electronic
documents.
Use of secure servers in the delivery of electronic legal gazettes
If there is a secure worklow in the production of the legal gazette, it may
be suficient to use secure servers with certiicates.
Use of workflow in the production of electronic legal gazettes
If the production of legal gazettes is decentralised and there is no secure
worklow in the production of the legal gazette, it may be necessary to establish
a secure worklow for the whole process or for the printing process.
01_2007_5222_txt_ML.indd 102 6-12-2007 15:13:51

WORKSHOP
Examples of different approaches
Methods Country Remarks
No speciication
Use of worklow
Use of secure servers and
certiicates
Use of electronic signatures
Belgium, United Kingdom,
Portugal, Norway
Austria, Denmark, france,
Portugal, Germany, finland
Estonia, france, Greece,
Portugal
Austria, france, Greece,
hungary, Slovenia
Legislative changes
No technical projects
concerning authenticity of
ofical journal
the scope of worklow
varies: only the publishing
or the whole lawmaking
process
mostly servers with secure
https protocol
Also open source products
Different signatures: server
signatures, XmL
signatures, XaDES
signatures, PDf signatures
2. ON thE KEy CONCEPtS
the working group has noticed that there is a plethora of concepts used in
the discussion on electronic publishing of legislation. to clarify the differences
between the key concepts, the working group has drawn up a list of key concepts
with short descriptions or deinitions.
the key concepts in the electronic publishing of legislation are:
• authenticity/authentication of electronic documents,
• chain of conidence,
• digital or electronic signature/advanced or qualiied electronic signature,
• electronic documents,
• oficial electronic versions.
Authenticity
Authenticity refers to the quality and credibility of an electronic document
(in this case the electronic act or the electronic legal gazette). It has something
in common with genuineness, legitimacy, undisputed credibility and believability.
Authentic implies being fully trustworthy in accordance with fact and
with regard to documents in law, and authenticity (Greek: αυθεντικός, from
‘authentes’ = ‘author’) is the truthfulness of origins, attributions, commitments,
sincerity and intentions, and not being a copy or forgery. As regards
electronic legal text, authenticity means that text published is provided by the
competent authority, that the text has not undergone any substantial
102 | 103
01_2007_5222_txt_ML.indd 103 6-12-2007 15:13:52

modiication and that the text read today is strictly identical to text irst released
perhaps 50 years ago.
webster’s 1913 dictionary deines authenticity as ‘the quality of being authentic
or of established authority for truth and correctness’. It also refers to
genuineness; the quality of being genuine or not corrupted from the original.
In the electronic world and in digital materials, authenticity means that the
digital material is what it purports to be. In the case of electronic documents,
such as electronic legal gazettes, it refers to the trustworthiness of the electronic
document as a document. the authentic version could be described as
the complete digital image of the paper version. In the case of originally digital
(‘born digital’) or digitised materials, it refers to the fact that whatever is being
cited is the same as it was when it was irst created, unless the accompanying
metadata (e.g. on the amendments or the consolidation of the law) indicate
any changes.
On 15 September 2001, the International Organisation for Standardisation
(ISO) published the standard ISO 15489-1, Information and documentation
— Records management. this standard, in Clause 7.2.2, deines authenticity:
‘An authentic record is one that can be proven (a) to be what it purports to be,
(b) to have been created or sent by the person purported to have created or
sent it, and (c) to have been created or sent at the time purported.
to ensure the authenticity or records, organisations should implement
and document policies and procedures which control the creation, receipt,
transmission, maintenance and disposition of records to ensure that records
creators are authorised and identiied and that records are protected against
unauthorised addition, deletion, alteration, use and concealment.’
Authentication can be deined as the process of verifying that a document
or message is authentic and that is has not been altered en route from the producer
of the document to the recipient(s). Authentication systems have become
an essential part of electronic commerce and e-government via the Internet.
Based on a range of encryption techniques, digital signature systems allow
organisations and individuals to electronically certify the authenticity of an
electronic document. Authentication has also another meaning — the authorisation
of a person to access an electronic system.
Chain of conidence
A chain of conidence (chaîne de confiance) is a reliable process for producing
electronic documents, which can be authentic and oficial. Usually the chain of
01_2007_5222_txt_ML.indd 104 6-12-2007 15:13:52

WORKSHOP
conidence uses certiication and digital signatures. It is important that there
are methods for the diagnosis and veriication of the integrity of the process
and of the different phases.
Digital or electronic signature
An electronic signature is used to authenticate the identity of the sender of
a message or of the signer of a document (e.g. act of parliament). It can also be
used to ensure the integrity of the original content of a document or message.
Additional beneits to the use of a digital signature are that it is easily transportable,
cannot be easily repudiated, cannot be imitated by someone else, and can
be automatically time stamped.
In the Uncitral model Law on Electronic Signatures (2001), electronic signature
means data in electronic form in, afixed to or logically associated with
a data message, which may be used to identify the signatory in relation to the
data message and to indicate the signatory’s approval of the information contained
in the data message.
Directive 1999/93/EC of the European Parliament and of the Council of 13
December 1999 on a Community framework for electronic signatures ( 2 ) deines
an electronic signature as: ‘data in electronic form which is attached to or
logically associated with other electronic data and which serves as a method of
authentication’. the directive addresses three forms of electronic signatures.
the irst one is the simplest form of ‘electronic signature’ and is given a wide
meaning. It serves to identify and authenticate data. It can be as simple as signing
an e-mail message with a person’s name or using a PIN code. to be a signature,
the authentication must relate to data and not be used as a method or
technology only for entity authentication.
Advanced and qualiied electronic signature
An advanced or secure electronic signature connects the signature more
closely with the signatory. A secure electronic signature is deined as ‘an electronic
signature that results from the application of a technology or process
prescribed by regulations’.
In Directive 1999/93/EC, the second form of electronic signature deined
is the ‘advanced electronic signature’. this form of signature has to meet the
requirements deined in Article 2(2) of the directive. In this article, advanced
( 2 ) Oficial Journal of the European Communities L 13, 19.1.2000, p. 12.
104 | 105
01_2007_5222_txt_ML.indd 105 6-12-2007 15:13:52

electronic signature is an electronic signature which meets the following requirements.
• It is uniquely linked to the signatory.
• It is capable of identifying the signatory.
• It is created using means that the signatory can maintain under his or her
sole control.
• It is linked to the data to which it relates in such a manner.
the directive is technology-neutral but in practice this deinition refers
mainly to electronic signatures based on a public key infrastructure (PKI). this
technology uses encryption technology to sign data, which requires a public
and a private key.
In addition, there is a third form of electronic signature mentioned in Article
5(1), which the directive did not give a term of its own but which for the
purposes of this report will be called ‘qualiied electronic signature’. this consists
of an advanced electronic signature based on a qualiied certiicate and
created by a secure-signature-creation device and needs to comply with the
requirements in Annexes I, II and III of the directive.
An example of the use of qualiied electronic signature is in the Austrian
database of authentic legislation (BGBL Autentisch), where there are four formats
available: htmL, PDf, word and authentic XmL text. In the Austrian
service, the digital signature can be viewed and veriied separately.
Electronic document
An electronic document is deined as ‘data that is recorded or stored on
any medium in or by a computer system or other similar device and that can
be read or perceived by a person or a computer system or other similar device.
It includes a display, printout or other output of that data.’
Oficial electronic version
Is a version of the law published in electronic form and recognised by the
national authorities as an oficial electronic document. the oficial electronic
version may have a similar status to the paper version of law. the status of the
electronic version can be based on a speciic act, administrative procedure or
other authorisation. the oficial electronic version can be equal to the paper
version of the legal gazette or (as in some countries) the electronic version can
be the only oficial version.
01_2007_5222_txt_ML.indd 106 6-12-2007 15:13:52

WORKSHOP
3. LEGISLAtIVE ISSUES CONCERNING LEGAL GAZEttES —
thE RELAtION Of thE ELECtRONIC VERSION tO thE PAPER
VERSION
3.1. The legal status of the electronic legal gazette
the member States of the European forum of Oficial Gazettes have different
approaches to the legal status of the electronic legal gazette.
THE TRADITIONAL APPROACH
the traditional approach to the deinition of the legal status has been that
the paper version is the only authentic and legally valid version of the legal
gazette. In addition, it has been possible to publish unoficial, unauthentic
electronic versions of the legal gazette, mostly for information purposes and
for easier access to the materials of the legal gazette. Usually these electronic
gazettes are available in htmL or PDf formats. Nowadays the traditional approach
is still valid in a large number of European countries.
THE BALANCED APPROACH
the balanced approach in the deinition of the legal status made progress
in the late 1990s in a number of countries. In the balanced approach, the paper
version and the electronic version have equal legal status. however, if there is
a difference in the content of the paper version and the electronic version, the
paper version is usually the only authentic one.
In France, the electronic publication of the authenticated oficial journal
has acquired a status equivalent with the paper version since 1 June 2004, and,
by coming out at the same time as the paper version, it constitutes the oficial
publication of legal acts. the electronic oficial journal is not identical with the
paper journal, due to the question of anonymising regulatory acts relating to
the status of individuals. Name changes or naturalisation are sensitive issues
and therefore notices or acts including information on name changes or naturalisation
are not published digitally.
Another example of this balanced approach is the Estonian act on the
legal gazette (Riigi Teataja) from 1999. the act (paragraph 1) states the following.
• the Riigi Teataja is the oficial publication of the Republic of Estonia. the
Riigi Teataja shall be published on paper as printed matter and electronically
on the Internet.
• Legislation, notices and other documents published in the Riigi Teataja
which is on paper and in electronic form have equal legal force, unless this
act provides otherwise.
106 | 107
01_2007_5222_txt_ML.indd 107 6-12-2007 15:13:52

NB: In Estonia the Riigi Teataja Act has been amended in such a way that
the electronic version is the only authentic one from 1 January 2007 onwards
and there will be only ive paper copies for archiving.
Another balanced and eficient approach to authenticity can be deined as
declaratory authenticity. In declaratory authenticity, the authenticity of electronic
versions is deined by either:
• national law, for example with the deinition that ‘the electronic version
has the same legal status as the paper version’ or the ‘the electronic version
is the only authentic version’; or
• declaration, tradition or administrative principle: the declaration can be
based on the competence and authority of the publisher or on general reliability
of the source. this is the case, for example, in the United Kingdom
and Norway.
THE INFORMATION SOCIETY APPROACH
the information society approach to the deinition of the legal status is a
fairly recent one. During the last four years there has been — in a number of
countries — new legislation on the paper and electronic versions of the legal
gazette. these new legislative acts deine the electronic version as the primary
authentic version. In comparison to the traditional approach, the situation is
upside down: it is possible to publish unoficial, unauthentic paper versions of
the legal gazette, mostly for information purposes and for the archiving of the
materials of the legal gazette. In some countries, the number of these paper
copies has been limited to between four and six copies.
Belgium was among the irst countries to reduce the number of printed
copies. the situation in Belgium is now as follows: distribution to the public
has been carried out exclusively online since 1 January 2003. Only ive paper
copies of each legal gazette are produced (for the ministry of Justice, Moniteur
belge, the National Library, National Archive, and one for microilming) for different
purposes: archiving, accessibility and continuity of the principle of publication.
the only way for individuals to consult the oficial journal is in PDf
format on the secure site of the federal ministry of Justice (www.moniteur.be).
A free-of-charge help desk for facilitating access to the online version was created
on 1 August 2005, and since 15 October 2005 almost every courthouse
has been obliged by law to make paper summaries of the legal gazette available
for the public.
In Austria, the publishing of the legal gazette was reformed from 1 January
2004. the legally binding Austrian federal Law Gazette, with electronic
01_2007_5222_txt_ML.indd 108 6-12-2007 15:13:52

WORKSHOP
signature, is only published in the Austrian Legal Information System. According
to the federal Act on the federal Law Gazette (Bundesgesetzblattgesetz
BLBlG, 2003: 100, paragraph 7) ‘legislation to be published in the federal Law
Gazette shall be available for access on the Internet at the address www.ris.
bka.gv.at. Each issue of the federal Law Gazette shall make reference to this
address.’ the unoficial paper copies still exist for archiving purposes (paragraph
8): ‘At least three back-up copies and four certiied printouts shall be
made of each document. One back-up copy and certiied printout each shall be
delivered to the Austrian National Archive and to the Austrian National Library
to be iled there. One certiied printout shall be forwarded to the Library of
Parliament.’
In Slovenia, a reform was carried out in 2005. In the Slovenian Oficial
Gazette Act of 2005, the new approach is clear (Articles 1 and 4): ‘the oficial
gazette shall be published in electronic form and in a printed version. the
electronic edition of the oficial gazette shall be safely electronically signed and
published on the website of the public company Uradni list Republike Slovenije.’
‘the electronic edition and printed version of the oficial gazette shall
be published on the same day. when both editions are not published on the
same day or do not contain the same text, the electronic edition shall apply.’
In Portugal, a new act on the oficial journal entered in force on 1 July
2006, changing the organisation, entry into force and the distribution of the
oficial journal. the oficial journal Diário da República was reduced from three
to two series.the irst series used to be divided in two parts, which then
were incorporated into one.the third series, the content of which dealt with
societies, was discontinued. the acts of establishing and modiication of societies,
which represented 80 % of the third series, were moved to another
website.
from 1 July 2006 the electronic version became the authentic oficial journal.
the acts only become oficial after their publication in the electronic journal
at the website www.dre.pt. the paper version of the oficial journal was
discontinued on 31 December 2006. It will be possible to have the paper version
of the irst series, but only for private societies and the general public, at
real prices and only for the purpose of information. the distribution of the
paper version to the government and to the local and regional administrations
is forbidden.
108 | 109
01_2007_5222_txt_ML.indd 109 6-12-2007 15:13:52

3.2. Speciic legal issues in electronic publishing
In the electronic publishing of legal gazettes, a number of speciic issues
have to be analysed.
3.2.1. SIMULTANEOUS PUBLISHING — IS THERE ANY NEED
TO REGULATE THE PUBLISHING ORDER OF THE PAPER
AND ELECTRONIC VERSION (WHICH IS PUBLISHED FIRST)?
In most European countries, the paper and electronic versions are published
simultaneously, although in practice the electronic version is naturally
available earlier than the paper version.
3.2.2. IS THE ENTRY INTO FORCE OF THE ACT DEPENDENT
ON THE PAPER VERSION OR THE ELECTRONIC VERSION
traditionally, the acts have entered into force on the day when the paper
version is available or some days later (‘one day after the publication date of
the legal gazette’ or ‘within ive days’, etc.). Electronic publishing has changed
this situation. today, in a number of countries (e.g. Austria and france), the act
enters into force on the day it is published in electronic form. According to the
Austrian legislation (BGBlG paragraph 11), ‘Unless therein or by law speciied
differently, publications in the federal Law Gazette whose contents are binding
shall be effective as of expiry of the day of release for access. Each issue of
the federal Law Gazette shall contain this date.’
Similar changes are being prepared in, for example, Estonia.
3.2.3. IS THERE ANY REFERENCE TO ELECTRONIC SIGNATURES
IN LEGISLATIVE ACTS IN THE LEGAL GAZETTE?
the use of electronic signatures in the authentication of legislative acts is
also relected in the legislation. In only a few countries is there direct reference
to the use of electronic signatures. the Austrian federal Act on the federal Law
Gazette (paragraph 8) contains the following points.
• Documents containing legislation to be published must have a format
warranting upward compatibility. they must have been produced through
a reliable process and carry an electronic signature.
• After having been provided with the signature, the documents must not
be modiied any more and nothing more shall be deleted after they been
released for access.
A similar reference to electronic signatures is found in the Slovenian legislation.
01_2007_5222_txt_ML.indd 110 6-12-2007 15:13:52

WORKSHOP
3.2.4. ARE THERE ACTS, DECREES OR SECONDARY LEGISLATION
WHICH ARE PUBLISHED ONLY IN ELECTRONIC FORM?
the lower cost of electronic publishing has contributed to the recent development
that in some countries a number of decrees or secondary legislation
are published only in electronic form. this kind of electronic-only publishing
has been used in france, finland and Slovenia. (In Slovenia, ‘Other instruments,
the publication of which is stipulated by law or other regulation, shall
be published only in the electronic edition unless otherwise provided by law
or other regulation.’)
In france, Decree No 2004-617 of 29 June 2004 states that certain texts are
to be published in digital form only. the texts in question are:
• regulatory acts, other than orders, which concern the administrative organisation
of the State, in particular decrees relating to the organisation of
central administrations, acts relating to the organisation of the decentralised
services of the State, and those delegating signature authority within
the State’s services and its public institutions;
• regulatory acts, other than orders, which concern public oficials and other
staff, judges and servicemen;
• regulatory acts, other than orders, which concern the State budget, in particular
decrees allocating, opening, cancelling or transferring appropriations,
those which concern support funds, account headings of the State
treasury and imprest accounts, and also budgetary and accounting directives;
• individual decisions taken by the minister for Economic Affairs in the ield
of competition;
• regulatory acts of independent administrative authorities and independent
public authorities having legal personality, other than those concerning
the general public.
furthermore, in france, Decree No 2004-459 of 28 may 2004, adopted
pursuant to Article 4 of the Order of 20 february 2004, speciies the individual
acts which may not be published in electronic form. the decrees in question
are those concerning:
• changes of name,
• acquisition, recovery, loss or forfeiture of french nationality,
• naturalisation,
• Gallicisation of forename or surname or assignment of forename.
110 | 111
01_2007_5222_txt_ML.indd 111 6-12-2007 15:13:52

3.2.5. IS THERE ANY FORCE MAJEURE CLAUSE IF THE ELECTRONIC
VERSION IS NOT AVAILABLE?
the modern view on electronic publishing of legislation is based on the
fact that the Internet works and no hacking disturbs the access to law. the possible
problems have been taken into account in the legislation of some forum
member States. In Austria, the federal Act on federal Law Gazette includes a
force majeure clause (in paragraph 7): ‘If and as long as making or keeping legislation
to be published in the federal Law Gazette available for access on the Internet
is not only temporarily impossible, such publication shall take place in a
different manner complying with the requirement of Article 49(3) BVG.’
4. USE Of ELECtRONIC SIGNAtURES
the electronic signature (as ‘data in electronic form which is attached to or
logically associated with other electronic data and which serves as a method of
authentication’) can be related to a person or to an institution (corporate signature,
server signature). today there are a large number of standards or de facto
standards for electronic signatures. the electronic signatures can be used as
server-based signatures or as signatures contained in the electronic documents.
XAdES — a European ETSI standard for electronic signatures
An electronic signature produced in accordance with the XAdES standard
provides evidence that can be processed to get conidence that some commitment
has been explicitly endorsed under a signature policy, at a given time, by
a signer under an identiier, e.g. a name or a pseudonym, and optionally a role.
the signature policy speciies the technical and procedural requirements on
signature creation and validation in order to meet a particular need.
XAdES extends the original XmL signature speciication with additional
syntax and processing necessary to satisfy the European directive on a community
framework for electronic signatures as well as other cases of use requiring
long-term validity. XAdES itself contains several modules that permit varying
levels of security such as non-repudiation with time stamps, certiication
data and certiication archives.
the European EtSI standard tS 101 733 for XAdES deines formats for
advanced electronic signatures that remain valid over long periods, are compliant
with the European directive and incorporate additional useful information
in common-use cases (such as indication of the commitment got by the signature
production).
01_2007_5222_txt_ML.indd 112 6-12-2007 15:13:52

WORKSHOP
Electronic signature of PDF documents
the Adobe Acrobat software supports a number of standards for electronic
signatures, especially the PKCS de facto standards.the best way to establish
that a PDf document is authentic and genuine is to check whether the
digital signatures (if any) within it are authentic. A PDf document can have
two kinds of digital signatures.
• A certiication signature, which can be applied by the document’s author:
Adobe Reader or Acrobat automatically checks the authenticity of this
signature when one opens the document, and then displays a window
that indicates whether the signature is valid (that is, authentic and current).
the certiication signature can be described as the ‘author’s digital signature’.
• A standard signature, which can be applied by anyone who has permission
to digitally sign the document: Adobe Reader or Acrobat can automatically
check the authenticity of standard signatures when one opens
the document, or one can check them manually from within the application.
(NB: Adobe Reader or Acrobat must have access to the Internet to
check digital signatures.)
CHECKING THE CERTIFICATION SIGNATURE FOR PDF DOCUMENTS
Immediately after a certiied Adobe PDf document is opened, Adobe
Reader or Acrobat automatically checks for unauthorised modiications to the
document and checks the authenticity of the certiication signature. the software
then opens a document status window that shows one of three results:
‘Certiication valid’, with a blue ribbon; ‘Validation of author not conirmed’,
with a blue question mark next to a person; or ‘Certiication invalid’, with a red
X. A valid result can provide strong assurance that the document has not been
modiied and that the document is genuine.
Challenges in the use of electronic signatures
there are several challenges in the use of electronic signatures. the irst
challenge is in the choice of the most suitable signature. there are several alternatives
and each of them has a different procedure for utilisation.
the second challenge is that electronic signatures are ageing. Increasing
computing power, the possibility of networking, and the progress of cryptography
contribute to the ‘weakening’ of electronic signatures, i.e. electronically
signed documents may lose their probative value over the years. Electronic
112 | 113
01_2007_5222_txt_ML.indd 113 6-12-2007 15:13:53

signatures usually have a limited time of validity. the challenge of renewal of
signatures will be signiicant in the coming years.
the third challenge is related to reform and data transfer. Electronic signatures
break when changing the document format. technological developments,
harmonisation attempts and also new legal guidelines cause changes of
user data and signature formats over the years. with electronically signed documents,
format changes are problematical, since changing the format breaks the
original signature. A similar problem arises during the digitalisation of paper
documents. If, for instance, a document signed by hand is digitised, the signature
loses its validity. the legal authenticity of the transformed document is at
least doubtful.
Practices
FRANCE
two types of electronic signatures are used in the chain of conidence.
XAdES is used in most cases with a higher level of authentication as a nonintrusive
signature, and PDf(#PKCS7) as an intrusive signature.
A time stamp for the XAdES texts can extend the certiicate period of
validity (beyond the initial two years), in accordance with the RfC 1305 Network
time Protocol and the speciic time-stamp protocol RfC 3161 Internet
X509 PKI time-Stamp Protocol (tSP). In france the software of choice for
managing the time stamp has been nCipher Appliance. It includes a new tool
to ‘re-sign’ the texts at the end of validity of the certiicate.
A Crypto Box is used to keep secure the private keys for the publication’s
signature. (these keys are not personal keys but bear the name of the oficial
gazette.)
AUSTRIA
An XmL-based digital signature has been used since 2004. the documents
leaving the secure worklow will be
signed electronically on an XmL
basis, using XmL-DSIG. It is a serverbased
electronic signature by the
federal Chancellery.
An example of the use of qualiied
electronic signature is in the
Austrian database of authentic legislation
(BGBL Autentisch), where four
01_2007_5222_txt_ML.indd 114 6-12-2007 15:13:54

WORKSHOP
formats are available: htmL, PDf, word and authentic XmL text. In the Austrian
service, the digital signature can be viewed and veriied separately.
GREECE
A PDf signature is used in the electronic version of the Greek legal gazette.
the PDf signature includes a link to the document status window or veriication
page, where it is possible to check the signature data and the validity
of the document.
114 | 115
01_2007_5222_txt_ML.indd 115 6-12-2007 15:13:55

In france, the user of the authentic version of the oficial journal can check
the validity of the electronic signature and view the certiicate.
5. wORKfLOw AND ChAIN Of CONfIDENCE
worklow solutions provide organisations with a means of automating
and streamlining content-centric processes and managing the life cycles of
those processes eficiently. worklow solutions are used to develop and optimise
multiple types of process automation applications — including processes
that involve both systems and people.
worklow is commonly associated with electronic processes for managing
documents. worklow handles approvals and prioritises the order in which
documents are presented. Authenticity by worklow is generally very demanding.
In the worklow, the integrity of the electronic act is secured in a work
process from the irst draft to the inal text published in the legal gazette. the
worklow can consist of capturing the data from source and transferring it
within one worklow or between separate worklows (ministries/parliament).
01_2007_5222_txt_ML.indd 116 6-12-2007 15:13:57

WORKSHOP
the chain of conidence is a reliable process for producing electronic documents
which can be authentic and oficial. Usually the chain of conidence
utilises certiication and digital signatures. It is important that there are
methods for the diagnosis and veriication of the integrity of the process and
of the different phases.
the chain of conidence may cover only a part of the production process,
such as the inal steps of preparing the publication from the materials received
from the ministries and the parliament. At the moment, there are very few
experiences of either worklow or chain of conidence.
Practices
AUSTRIA
the e-Recht (e-law) project has prepared a separate worklow for the ministries
and for the parliament. the basic ideas were:
• to replace printed legal texts by digitally signed electronic documents;
• to provide an electronic worklow for producing legal texts (e.g. laws,
regulations, announcements, treaties);
• to enable oficial publication of the Austrian federal Law Gazette on the
Internet.
the legislative process in Austria has been reformulated in the following
way.
• A draft of a bill is prepared by a ministry.
• the ministry sends the draft bill for internal consultation (expert’s opinion) to
different interest groups (e.g. trade unions, chamber of commerce).
• the draft bill can also be put into the RIS legal information system.
• A decision on it is taken at the Council of ministers (which is a weekly
meeting of the Austrian federal ministers).
• the draft bill becomes a government bill and has to be put into RIS.
• the government bill is transferred from the Government to the Parliament,
which runs an independent system.
• the government bill is discussed by the Parliament.
• the decision of the Parliament is transferred back to the federal Chancellery.
• the bill is signed by the federal President and countersigned by the federal
Chancellor on paper.
• A server-based electronic signature is provided by the federal Chancellery.
116 | 117
01_2007_5222_txt_ML.indd 117 6-12-2007 15:13:57

• there is oficial publication of the authentic version of the federal Law Gazette
in the public RIS information system (http://ris1.bka.gv.at/authentic/
index.aspx).
In the drafting of draft bills, documents are written in microsoft word
supported by dedicated macros. All documents are structured by legal categories
(formats). the correct use of the special templates is necessary for the
conversion of the document to XmL. Important functions of the templates are
automatic format recognition (interactive/quick), e-law validation (protocol),
creation of a table of contents for the law in the draft bill, and comparison of
text modules.
In Austria, a speciic mOA (module for online applications, with modern
security technology, consisting of server modules and security modules for
signature creation, signature veriication, identiication, delivery, etc.) is available
free of charge (except for third-party libraries). there are different types of
worklow for laws, regulations and treaties.
FRANCE
A chain of conidence was initially used only in the publishing process.
the chain of conidence was organised by the oficial journal directorate for
the purpose of authenticating the electronic version of the oficial journal. the
chain of conidence was based on the probative value of electronic evidence
being given the same status as the traditional paper document. to back up this
principle, the legislator allowed two conditions on the principle of equality:
‘that it should be possible for the person from whom it emanates to be duly
identiied and that it should be kept under conditions that will ensure integrity.’
these conditions were incorporated into the chain of conidence.
In france, a government worklow system called SOLON has been established
(SOLON = Système d’organisation en ligne des opérations normatives).
01_2007_5222_txt_ML.indd 118 6-12-2007 15:13:58

WORKSHOP
the object of SOLON is to manage the low of primary legislation between
the ministries and the parliament and the low of secondary legislation.
SOLON ensures the validation low (with corrections, if needed) of a text that
has already been written until publication.
this project is managed by the General Secretary of the Government, carried
out on a technical level by SAP and hosted by a private company at equipment
level. the oficial journal is part of the technical team working on the
project. SOLON is the front end of the oficial journal production.
the basis of the system is the SAP software but SOLON requires some
speciic functions to be done by parameterisation and some extra developments
to personalise the tool. the preliminary work, necessary to get texts
written corresponding to primary legislation or secondary legislation, is not
part of SOLON. the data capture is done with microsoft word before downloading
the content into Solon and the users must use the style sheets designed
for the different type of acts produced as e-legislation. the ‘driving open ofice’
in version number 2 permits an XmL translation based on the use of some
predeined style sheets.
the SOLON system concerns initially secondary legislation, managed by
ministries. the primary legislation will use SOLON as soon as the secondary
legislation is in full production in 2007. three ministries are connected, plus
the Oficial Secretary of the Government and the oficial journal. SOLON’s
access is at ministry level — cabinet ofices, State Council, Assemblies and
Prime minister’s Secretaries, (contributions) and inally the oficial journal (reception).
SOLON is used by about 500 people, and is in reality an elaboration
of the administrative processing of texts, for the whole validation low.
SOLON uses the private network of administration called ADER.
PORTUGAL
A system calle RedeLex — the network for the electronic legislative procedure
— is being established.
the RedeLex project allows the interconnection, in a safe and private network,
of some entities and agencies of sovereignty that participate in the electronic
legislative procedure (the Government, the Presidency of the Republic,
the Assembly of the Republic and the Constitutional Court). It establishes a
connection between this network and the General Secretariat of the Presidency
of the Council of ministers and the national printing house. RedeLex will use
qualiied digital signatures and safe data transmission.
118 | 119
01_2007_5222_txt_ML.indd 119 6-12-2007 15:13:58

FINLAND
A system called PtJ (Government Decision-making System) is the network
for electronic legislative procedures. the irst PtJ was established in 1995 and
the new PtJ2 was introduced in 2005, based on Documentum software. the
draft laws are transferred within the system from the ministries to government
sessions and on to the Parliament in structured XmL format. the tool for the
actual drafting is microsoft word, with tailor-made templates and macros.
6. thE USE Of SECURE SERVERS AND CERtIfICAtES IN thE
DELIVERy Of ELECtRONIC LEGAL GAZEttES
the use of secure servers or secure protocol in the delivery of electronic
legal gazettes ensures the reliability of any electronic document source and the
transfer of data from the server, e.g. from the server of the legal gazette. A secure
server provides secure connections and the data in the in-transit process
between the user and the server is encrypted.
SSL (secure socket layer) server authentication allows users to conirm a
web server’s identity. SSL-enabled client software, such as a web browser, can
automatically check that a server’s certiicate and public ID are valid and have
been issued by a certiicate authority (CA) — such as VeriSign, thawte or
Geotrust — listed in the client software’s list of trusted CAs. SSL server authentication
is essential for secure e-commerce or e-government transactions.
An encrypted SSL connection requires all information sent between a client
and a server to be encrypted by the sending software and decrypted by the
receiving software, protecting private information from interception over the
Internet. In addition, all data sent over an encrypted SSL connection is protected
with a mechanism for detecting tampering — that is, for automatically
determining whether the data has been altered in transit.
Secure sessions and server IDs work in the following way.
1. A user contacts the site of the legal gazette (‘LG’) and accesses a secured
URL — a page secured by a Server ID (indicated by a URL that begins with
‘https:’ instead of just ‘http:’ or by a message from the browser).
2. the LG server responds, automatically sending the user the digital certiicate
of the LG site, which authenticates the site.
3. the web browser of the user generates a unique ‘session key’ to encrypt
all communications with the LG site.
4. the browser encrypts the session key itself with the site’s public key so
that only the site can read the session key.
01_2007_5222_txt_ML.indd 120 6-12-2007 15:13:58

WORKSHOP
5. A secure session is now established. It all takes only seconds and requires
no action by the user. Depending on the browser, the user may see a key
icon becoming whole or a padlock closing, indicating that the session is
secure.
A secure server is a web server that supports any of the major security
protocols, like SSL, which encrypt and decrypt messages to protect them
against third-party tampering. major security protocols include SSL, ShttP
(Secure httP), PCt, and IPSec (Internet Protocol Security).
A secure protocol (e.g. httPS) can be used to protect the transfer of data
from a secure server, with a security protocol such as SSL, tLS or PCt. Usually
the use of certiicates (by thawte, VeriSign, multicert, etc.) is recommended,
in order to verify that the address of the web server actually belongs to the
publisher of the legal gazette.
the SSL protocol runs above tCP/IP and below higher-level protocols
such as httP or ImAP. It uses tCP/IP on behalf of the higher-level protocols,
and allows an SSL-enabled server to authenticate itself to an SSL-enabled client,
the client to authenticate itself to the server, and both machines to establish
an encrypted connection.
the client uses the public key from the certiication authority’s certiicate
to validate the CA’s digital signature on the server certiicate being presented.
If the information in the server certiicate has changed since it was signed by
the CA, or if the CA certiicate’s public key does not correspond to the private
key used by the CA to sign the server certiicate, the client will not authenticate
the server’s identity. At this point, the client has determined that the server
certiicate is valid.
120 | 121
01_2007_5222_txt_ML.indd 121 6-12-2007 15:13:58

the next step conirms that the server is actually located at the same network
address speciied by the domain name in the server certiicate. Users
must perform this step and must refuse to authenticate the server or establish
a connection if the domain names do not match. If the server’s actual domain
name matches the domain name in the server certiicate, the client goes on to
next step, where the server is authenticated.
the SSL protocol supports the use of a variety of different cryptographic algorithms,
or ciphers, for use in operations such as authenticating the server and client
to each other and transmitting certiicates. well-known algorithms include:
• DES (Data Encryption Standard): an encryption algorithm used by the US
government,
• DSA (Digital Signature Algorithm): part of the digital authentication standard,
• RSA: a public-key algorithm for both encryption and authentication, developed
by Rivest, Shamir and Adleman.
Among the open-source products, the modSSL module provides strong
cryptography for the Apache 1.3 web server via the Secure Sockets Layer (SSL
v2/v3) and transport Layer Security (tLS v1) protocols by the help of the
open-source SSL/tLS toolkit OpenSSL. Additional information can be found
at www.modssl.org. further information on OpenSSL can found at http://
www.openssl.org/about/.
Practices
FRANCE
A secure server with a certiicate is used in the Journal officiel.
01_2007_5222_txt_ML.indd 122 6-12-2007 15:13:59

WORKSHOP
ESTONIA
A certiicate-based httPS web server has been established for the legal
gazette. It incorporates a thawte certiicate, which guarantees validity.
PORTUGAL
A certiicate-based httPS web server has been established for the legal
gazette. It incorporates a multicert certiicate guaranteeing validity, which
means that a root authority has taken steps to verify that the web address actually
belongs to the publisher of the legal gazette, INCm.
7. OthER mEASURES Of AUthENtICAtION
7.1. Software for online authentication
In addition to these different methods of authentication of electronic legal
gazettes, a number of other measures can be used. In Hungary, the user of an
electronic legal gazette has the possibility to check online the authenticity of
the document using speciic software. the hungarian publisher of the legal
gazette, magyar hivatalos Közlönykiadó, has developed a separate technological
process to the electronic version of legal texts and oficial journals, based on
an ShA-512 algorithm. It determines whether electronic versions of oficial
journals that were earlier downloaded or were purchased are oficial or not. the
monitoring program and the instructions for use can be downloaded from the
‘hitelesség’ (= authentiication) link of the website www.magyarkozlony.hu.
the authentication program (Auth) asks for the location and name of the
oficial journal ile and sends the hash code made from the user’s ile to the
server of mhK, where — after identiication — the answer yes or no is given.
however, experience shows that very few users have actually used the possibility
to conirm the authenticity of the electronic legal gazette in hungary.
Authentication with speciic software may be dificult or even impossible on
certain workstations, if the user is not allowed to install the software.
7.2. Use of time stamps in the resigning of electronic
signatures
the validity of electronically signed legal gazettes or other documents will
become a serious challenge in the near future, since electronic signatures are
usually only valid for a limited period, such as two or three years.
In france, the Journal officiel has experience in the re-signing of electronic
documents. the texts of legal gazettes will have to be re-signed after the
122 | 123
01_2007_5222_txt_ML.indd 123 6-12-2007 15:13:59

expiry of their certiicates (except with time stamping) or algorithms (RSA 1024,
RSA 2048, ShA 1). the choice of electronic signature is very important to minimise
the tasks to re-sign, but it is important to check compatibility with the tools
used (navigators, Adobe Reader, Class Java, etc.). the irst re-signing operation
in france was carried out at the ‘Direction des journaux oficiels’ in June 2006: all
texts from 2 June 2005 onwards were re-signed (nearly 60 000 iles altogether).
It took a week to conduct the ‘re-signature operation’ in accordance with good
practice. the re-signing was conducted by a signatory with a technician and followed
a protocol close to the original task of signature.
Concerning time-stamp standards, a new OpentSA project has been started.
Its aim is to develop an RfC 3161-compliant, stable, secure, open-source
and free time-stamping authority client and server application. the following
deliverables have already been produced.
• A time-stamp patch for OpenSSL: the time-stamp request creation, response
generation and response veriication functionality is implemented
as an extension to the latest stable version of OpenSSL. this patch adds a
new tS command to OpenSSL with which the time-stamping operations
can be carried out. this patch and the time-stamp client have been merged
into the oficial version of OpenSSL and will be available from OpenSSL-
0.9.9 onwards.
• A time-stamp client: A simple command-line-driven tSA client that can be
used for creating and sending time-stamp requests over httP or httPS
to a tSA and for receiving and verifying the responses. the utility is distributed
with the OpenSSL time-stamp patch.
• A time-stamp module for Apache: this package is an extension module for
the latest stable version of the Apache httP server. Using the functionality
of the OpenSSL time-stamp patch, this module functions as an RfC 3161-
compliant time-stamp server over httP and httPS transport protocols.
Issued time-stamp tokens can be stored in a mySqL or fireBird database.
further information can be found at www.opentsa.org.
8. GOOD PRACtICES IN thE AUthENtICAtION
Of LEGAL GAZEttES: A PRELImINARy INVENtORy
8.1. General principles
CUTTING DOWN THE NUMBER OF PAPER COPIES
It is fairly easy to cut down the number of paper copies if the number of
subscribers to the oficial gazette is, for example, between 500 and 2 000. Nev-
01_2007_5222_txt_ML.indd 124 6-12-2007 15:14:00

WORKSHOP
ertheless, some assistance should be given to citizens who do not have access
to electronic media. Possible solutions are to start a help desk (as in Belgium)
or to support access to paper documents at local libraries (as in Estonia) or at
district courts (Belgium). for archiving, a few copies should be made and
handed over to the national library and national archives.
DEFINING THE LEVEL OF INFORMATION SECURITY AND
AUTHENTICITY — HOW MUCH AUTHENTICITY IS ENOUGH?
with electronic copies being utilised to a growing extent, the risk that the
electronic copies may not be authentic should be kept in mind. for the majority
of users, the authenticated electronic versions of acts are not necessarily the
most usable ones. for utilising and distributing the electronic versions, the
authentic one may not be best one either, as studies in Austria and france have
shown. In Austria, 51 % currently use the PDf version of the legal gazette,
36 % use the htmL version, 11 % use the word version and only 2.3 % use
the authentic version, which incorporates an electronic signature.
there are broadly similar results from Portugal, where the electronic version
has been authentic since July 2006. It is accessible via a secure system,
using httPS, and a non-secure system, with traditional httP protocol. Currently
only about 7 % use the secure system.
USING ELECTRONIC SIGNATURES
Nowadays electronic signatures are used with electronic legal gazettes in
only a small number of countries (Austria, france, Greece and Slovenia). the
signature keys develop quickly and will have to be changed soon in many
countries, for which a solution has to be found.
One practical problem with electronic signatures is the limited period of
validity. As a result long-term archiving will become a challenge in the near
future.
8.2. Steps towards authenticity — A checklist
1. CHECK THE AVAILABILITY OF PDF
make PDf iles available if the legal gazette is not yet available in PDf
format.
(a) PDf is a standard format in the information market (easy to use, easy to
transfer).
(b) XmL is another option.
124 | 125
01_2007_5222_txt_ML.indd 125 6-12-2007 15:14:00

2. CHECK THE PRODUCTION CHAIN Of thE LEGAL GAZEttE
make it as secure as possible, by including:
(a) checkpoints in the legislative process,
(b) checkpoints in the printing house,
(c) possible use of electronic signatures and certiicates.
3. CHECK THE LEGISLATION
the legislative norms on the publishing of the electronic version of the
legal gazette need to be prepared carefully, taking into account:
(a) the legal status of the electronic version: the primacy of the electronic version
or balanced relation between the paper and electronic versions,
(b) the entry into force of the electronic version and paper version,
(c) the possible need to make reference to the use of electronic signatures in
the electronic version,
(d) the possibility to publish electronic-only acts/decrees/secondary legislation,
(e) the need to publish consolidated acts in electronic form,
(f) the access to law: is the electronic version free of charge?,
(g) a force majeure situation: what is to be done if the electronic publishing is
not working?
4. CHECK THE USE OF SECURE SERVERS AND CERTIFICATES
(a) firstly, check and secure the physical protection of document databases
and access control to the original databases (irewall, etc.).
(b) Establish a secure server e.g. with httPS protocol. SSL certiicates activate
the secure padlock using httPS and assure the visitors that data sent
via the Internet are secured by using data encryption.
(c) Utilise secure servers with open architecture and lower costs. for example,
Plone is a ready-to-run content management system that is built on the
free Zope application server. Zope is an open-source web application server,
featuring a transactional object database which can also store dynamic
htmL templates, scripts, a search engine, and relational database
(RDBmS) connections and code.
(d) Utilise open-source server software applications, e.g. Apache SSL.
5. CHECK THE USE OF ELECTRONIC SIGNATURES
If the use of electronic signatures is preferred, choose the most suitable
type of electronic signature for your purposes.
(a) Is it necessary to use electronic signatures (e.g. when you already have a
reliable production process and a secure server)?
01_2007_5222_txt_ML.indd 126 6-12-2007 15:14:00

WORKSHOP
(b) If the system is XmL-based, use XmL-DSIG or XAdES or OpenXAdES.
(c) If the system is PDf-based, use PDf-digital signature (based on PCKS#7
standard).
(d) try to keep the utilisation of electronic signatures lexible and at the most
suitable level.
(e) If archiving is necessary, consider the use of time stamps (easiest with
XAdES or general time-stamp protocol standard, RfC 3161 Internet X509
PKI time-Stamp Protocol)
6. CHECK WHICH POSSIBILITIES TO USE FOR WORKFLOW AND
CHAIN OF CONFIDENCE
(a) try to minimise the number of different worklows.
(b) Start with a limited scope of chain of conidence.
(c) Explore the possibilities of using generic elements (e.g. Austrian generic
elements: modules for Online Applications) for the signature veriication,
identiication and delivery of documents.
(d) Consider the possibilities to use standard software and generic wordprocessing
applications and standards (e.g. DigiDoc).
(e) Speciications have to be well-deined and the steps of worklow should
be easy to change, if needed.
(f) Software in the worklow should be easy to use and easy to understand.
(g) Determine which way to progress from word keyboarding with style sheets
to a well-structured XmL ile (with a speciic XmL editor for drafting).
(h) Steps of the worklow should be easy to change.
7. KEEP THE USERS INFORMED ABOUT YOUR AUTHENTICITY
POLICY
tell the users what is the level of security and authenticity, use fAq pages and
updated documents on the authentication methods used in your legal gazette.
8. PAY ATTENTION TO THE STANDARDS AND METHODS USED BY
THE CONTRACTORS AND SUBCONTRACTORS
(a) Use the Common Criteria standards.
(b) Observe the other existing standards (in electronic document management,
metadata, long-term preservation of electronic documents, electronic
signatures, time stamping, secure servers, etc.).
(c) Utilise the open-source software and open-source solutions.
9. CHECK PROPORTIONALITY
(a) Evaluate the costs and beneits of the different techniques and methods.
126 | 127
01_2007_5222_txt_ML.indd 127 6-12-2007 15:14:00

(b) Consult the users of your services and ind out the practical needs of the
users of legal gazettes.
10. CARRY OUT RISK ANALYSIS
(a) make a SwOt analysis of the different techniques and methods.
(b) Do not become too enthusiastic about technical solutions.
01_2007_5222_txt_ML.indd 128 6-12-2007 15:14:00

COHERENCE OF TERMINOLOGY
AND SEARCH FUNCTIONS ( 1 )
firstly, I would like to thank the Publications Ofice very warmly for its kind
invitation to attend this very interesting conference and for giving me the
chance to present my project, called ‘Syllabus’.
more than a project, I will be trying to present a tool, a piece of software
which we have produced and which is related to applied ontologies and European
law in the ield of consumer law.
I am a comparativist and comparative law has proved to be a very good
ally in several projects of harmonisation of European private law. I think the
main reason is that comparativists do try to build bridges between different
legal systems.
this is the background of the project I am going to present to you. I was
working on a functional comparison in the ield of trust law between English
law and Continental law. As you know perfectly well, trust law as such does
not exist. my research group got in contact with colleagues from the University
of turin who were working on artiicial intelligence and developing applied
ontologies in the ield of law. we decided to combine the two research
groups — our competence in comparative law and their expertise in artiicial
intelligence — and we have come up with a product
as a result of four years of work.
the full name of the product — and indeed the
website, which is already available on the Internet —
is the ‘Legal taxonomy syllabus’ (http://normas.
di.unito.it/syllabus).
this Syllabus is the result of four years of research
which was sponsored by the European Commission
as part of an ‘Improving human potential’
network. It mainly concerned funding general grants
for three years for post-doctoral students from seven
GIANMARIA AJANI
Professor of the Law Faculty
of the University of Turin,
Italy
( 1 ) transcription of the oral presentation.
128 | 129
01_2007_5222_txt_ML.indd 129 6-12-2007 15:14:00

universities: turin (Italy), which was the lead institution, Barcelona (Spain),
Lyon 3 (france), münster (Germany), Nijmegen (the Netherlands), Oxford (the
United Kingdom) and warsaw (Poland). within this framework, we have received
competent young post-doctoral scholars from these universities in turin,
and we asked them to come up with a research programme on legal terminology.
the background of this study is the following. we were not satisied with
just translating words. we were not satisied, because you can easily translate
words without really catching the meaning of legal terms. you can translate
the German term Besitz using the Suisse term Besitz. Same language, same
name. But the institutions are different, because the meaning of the German
term Besitz — position — is not really the same in terms of the content of the
rule. the same applies to many other very well-known examples.
So we tried to address this issue, not in the sense of having a dictionary
— a multilingual dictionary — but in the sense of having a semantic tool.
‘Semantic tool’ means that we wanted to take into account the context in
which terms are used and to see how these terms were deined in the different
levels of European legislation, European case-law, national implementing acts
and national case-law. In order to make this task a bit more challenging, we
also added some doctrinal notes and comments.
Research on legal terminology is something different from research on
legal translation. It is also a priority identiied by the European Commission. In
its communication of 2003 on ‘Updating and simplifying the Community acquis’
( 2 ), the Commission indicates the need for ‘rewriting legal texts to render
them more coherent and understandable’ and for overcoming ‘potential legal
uncertainty resulting from inconsistent deinitions or terminology’. It is therefore
the Commission itself which recognises that there is a problem in terminology
regarding the acquis.
however, regardless of the way terminology in which is used, the deinition
of terminology itself is not coherent. Even in Commission documents the
word ‘terminology’ has different meanings: sometimes, it is meant as a kind of
substitute for searching common principles, sometimes it is used as a syno-
( 2 ) Communication from the Commission to the Council, the European Parliament, the
European Economic and Social Committee and the Committee of the Regions —
Updating and simplifying the Community acquis, COm(2003) 71 inal.
01_2007_5222_txt_ML.indd 130 6-12-2007 15:14:00

WORKSHOP
nym for legal categories, and at other times it is used simply as a synonym for
‘words’.
terminology is actually about something else. It is work on groups of legal
terms put into context. thus, the reason why we called our tool the ‘Legal
taxonomy syllabus’, with an emphasis on terminology, is that we wanted to
place the deinitions of the terms into context — context being the normative
background.
from this very moment, the perspective clearly appears — ending up with
a semantic tool, a research tool which enables the user to get the meaning and,
eventually, the translation of different related words, linked together. this is in
fact the same way Google works today.
we have decided to limit our research to consumer law for two simple
reasons. One is that there is a bulk of easily identiiable sets of rules: the directives
on consumer law as quoted in the Green Paper on European Union Consumer
Protection (COm(2001) 531 inal), together with a couple of important
directives which were added in 2003 and 2005. this made up a distinctive
corpus which was narrow enough to permit this exercise, but also general
enough to give sense to it, taking into account that consumer law today is still
the core of contract law within the European Union. the corpus is therefore
composed of the directives identiied in the 2001 Green Paper plus the 2005
directive on distance marketing of consumer inancial services ( 3 ).
we tried — and this is a very ambitious statement — to make this tool useful
for several groups of users, and I would say mainly to all those having an
interest in getting access to the meaning of consumer law in different languages:
lawyers, translators, legislators and scholars. when we mention legislators, we
refer particularly to the common framework of reference which was set by the
Commission in 2003. the idea is that the common framework of reference
should lead the European legislator to a coherent, restated set of regulations
from the acquis, which should enable better law-making and better regulation.
Entering now the structure of the Syllabus as such, the irst thing to say is
that it comprises 90 terms identiied within consumer law directives. these
terms are explored following a two-step approach. Step 1 gives the horizontal
perspective: it maps the occurrence of the selected terms in EU law and identi-
( 3 ) Directive 2002/65/EC of the European Parliament and of the Council of 23 September
2002 concerning the distance marketing of consumer inancial services and amending
Council Directive 90/619/EEC and Directives 97/7/EC and 98/27/EC, OJ L 271,
9.10.2002, pp. 16–24.
130 | 131
01_2007_5222_txt_ML.indd 131 6-12-2007 15:14:01

ies terminological variants. the project takes into account ive different legal
orders: the french, the Spanish, the Italian, the German and the British. the
ive boxes related to them are linked in the sense that the user can jump from
one legal system to another, and from one directive or one link to another, surfing
through the software. Step 2 provides the vertical perspective, comparing
the term used in EU law to that used in the national law of member States. the
user can, for instance, go from the European directive which has been written
in Italian into the Italian national piece of implementing legislation. the same
is valid for the french, Spanish, German and British legal systems.
In other words, the user can map the occurrence of the words within EU
law (horizontal perspective) and between EU law and the national level (vertical
perspective).
"
Step 1: The horizontal
perspective
Several EU directives
refer to the genral
principle of good faith
“Good faith”
Unfair contract terms directive (93/13/EEC)
Distance selling directive (97/7/EC)
Directive on distant marketing of inancial services
(2002/65/EC)
Unfair commercial practices directive (2005/29/EC)
Let us take the horizontal perspective irst. If, for instance, you compare
Directive 97/7/EC on distance sales ( 4 ) with Directive 2002/65/EC on the distance
marketing of consumer inancial services, you ind that the term ‘good
faith’ in the English texts is always expressed in the same way, using the same
words. the same occurs for the Spanish term buena fe. however, if you look
for the french version of the same directives you realise that what was expressed
as ‘good faith’ in English is translated into two different terms (loyauté
and bonne foi). the same happens as regards the Italian (lealtà and buona fede)
and the German (Lauterkeit and Treu und Glauben) versions.
( 4 ) Directive 97/7/EC of the European Parliament and of the Council of 20 may 1997 on
the protection of consumers in respect of distance, OJ L 144, 4.6.1997, pp. 19–27.
01_2007_5222_txt_ML.indd 132 6-12-2007 15:14:01

WORKSHOP
Article 4(2) of Directive
977/EC
EN “good faith”
ES “buena fe”
fR “loyauté”
It “lealtà”
DE “Lauterkeit”
“Good faith”
Article 3(2) of Directive
2002/65/EC
EN “good faith”
ES “buena fe”
fR “bonne foi”
It “buona fede”
DE “trau und Glauben”
Article 2(h) of Directive
2005/29/EC
Deinition of “diligenza
profesional”: rispetto a
pratiche di mercato oneste
e/o al principio generale
della bone fede nel settore
di attività del profesionista,
il normale grado della
speciale competenza
e attenzione che
reagioevolmente si possono
presumere essere esercitate
da un professionista nei
confronti dei consomatori
!
Step 1: The horizontal
perspective
Terminological
variants can be
identiied among
various languages
If you now consider the vertical perspective, i.e. going from the European
to the national level, you will see that the Italian word lealtà used in Directive
97/7/EC has been converted into buona fede e lealtà and the term buona fede from
Directive 2002/65/EC was changed into correttezza e buona fede. I took the Italian
example because I was more familiar with it but you have the same phenomenon
in German or french, or in other implementing measures.
EU
Article 4(2) of Directive
977/EC
“lealtà”
“Good faith”
Article 3(2) of Directive
2002/65/EC
“buona fede”
/ /
“lealtà”
“correttezza e buona fede”
Article 2(h) of Directive
2005/29/EC
“Le attività commerciali sono
improntate al rispetto dei
principi di buona fede, di
correttezza e di lealtà […]”
!
Step 2: The vertical
perspective
Again, as a result
of transposition,
terminilogical
fragmentation is
increased
It
Decr. Leg 185/1999
Decr. Leg 206/2005
Decr. Leg 190/2005 Decr. Leg 206/2005
So, is that important? Is that problematic? In a way it is, because we work
with words. when we hear the call for a more coherent European law, we have
to understand that it is not only European law — the acquis – which is at stake.
132 | 133
01_2007_5222_txt_ML.indd 133 6-12-2007 15:14:01

the relationship between the acquis and the national level should also be transformed
into something more coherent. the constructive support for this kind
of exercise is that it shows where minor interventions on legal text could enhance
coherence.
the second aspect of the problem is related to the concept of terminology
or taxonomy. It is not only a matter of using different words meaning the same
thing. Sometimes, when you move to different words you open a different
box and a taxonomy which differs from one term to another. this is very clear
in some cases related to the difference between obligations based on torts and
contractual obligations, where you can simply change some words and jump
from one sector to another and from one taxonomy to another taxonomy.
the website is already online. Access is free, so there is no password. the
homepage offers a short description of how the project was initiated, and information
on those who have contributed to it. the software is a new and
original tool, which was tailored on purpose by the informatics department at
the University of turin.
Once in the database, we can choose the Italian or the English interface.
we have the choice between ive languages and three levels: European and
national ‘all’, or only European or national. there is a box for ‘term name’. the
full list of terms, totalling about 100, is the corpus that has been worked out
from the set of directives on consumer law. A full list can be obtained by typing
an asterisk (*). the whole list represents 55 pages, in alphabetical order, in
all ive languages.
A test can be carried out, for example, on the term ‘offer’, with reference to
the English term ‘offer’ used in Directive 85/577/EEC. the full text of the actual
article of the directive where the term ‘offer’ has been deined and regulated can
be checked online. the site also gives associated information. finally the user
can compare the term ‘offer’ in other linguistic versions by clicking on ‘Show
other linguistic versions’ to obtain the term in English, Spanish, french, Italian
and German. we ind that ‘offer’ in German is Angebot. however, if we go to the
right-hand column (‘Associations’), we ind that Angebot generates two different
terms in the German implementing measures: Antrag and Angebot.
we are now at the national level, as we have jumped from the English term
‘offer’ used at European level to its synonym in German, and then moved into
the vertical level. the full text of the articles concerned can also be consulted
within the Syllabus. So the text of paragraph 145 BGB, which transposed the
ruling of the directive into German law, is available.
01_2007_5222_txt_ML.indd 134 6-12-2007 15:14:01

WORKSHOP
the tool could prove to be very useful for practicing lawyers and for
judges, because it lets the user shift very easily from one system of national
law to another.
the system works not only on single terms, but also on expressions, as
the example of ‘good faith’ showed. By clicking successively on ‘European’,
and then ‘Show other linguistic versions’, the user obtains the equivalent of
‘good faith’ in all ive languages and sees that there are several french equivalent
terms: loyauté, bonne foi, en traitant de façon loyale et équitable: clicking on ‘Association’
reveals the term used for national implementation: bonne exécution des
obligations.
Let me close this presentation with a word about ontology. the ontology
related to this programme has been elaborated by colleagues together with
computer scientists. It permits the development of an expert system which
will make the dictionaries within the Syllabus able to communicate with
others using analogue ontologies.
finally, the system allows for statistics. It gives a set of European terms
with distinctive transpositions in national law. It consists of a list of terms
used at European level and the corresponding — different — terms used at
national level. It illustrates how the language of European law differs from
the language of national law, even though it is the same, common language.
As an example, European law says Angebot for ‘offer’, but German national
law uses Antrag. the list is rather long and one could speculate about
whether this was done on purpose or whether there were simply mistakes.
when talking about access to data and databases using searches on terms,
problems might arise as these systems are rather rigid when it comes to recognising
or not recognising words. for instance, if the user searches for Angebot
in the national legislative level, he will not ind anything as there is nothing like
Angebot: German national law uses Antrag. thus the only way to perform fruitful
searches is to apply semantic tools. Semantic tools enable the user to capture
both Angebot and Antrag.
General statistics concerning the number of terms already worked out and
synonymies offer lists of terms used within the same context. this is where
the real connection between ‘ontology’ and Syllabus appears. Each entry provides
a single link in the ontology, which refers to all concepts.
As mentioned before, the project team tried to make the tool useful for
several groups of users. there is a level for ‘general users’ as well as a specialist
level: an ontologist can work out ontologies and can have access to the
134 | 135
01_2007_5222_txt_ML.indd 135 6-12-2007 15:14:01

ontological map. the idea is to make the tool available for different types of
users without necessarily making complex information available to common
users.
the project team welcomes all possible comments and reactions of users
experimenting with the Syllabus online!
01_2007_5222_txt_ML.indd 136 6-12-2007 15:14:01

EUR-LEX: FROM DATA STRUCTURES
TO LEGAL ONTOLOGIES
INtRODUCtION
Lawyers as knowledge workers have to cope with a tremendous load of
information. In the legal domain, almost all available information is stored as
text, most of the time in relatively unstructured forms (Stranieri and Zeleznikow,
2005). As legal work consists in solving legal problems, consultation of various
texts is a prerequisite of legal work. this legal research can be outsourced to
paralegals but at the very end highly qualiied lawyers have to reine the often
high quantity of search results in order to get a corpus of relevant legal texts.
Given the high costs and time constraints, improvements of existing legal
databases in order to reduce this process of reining search results would be
highly welcome to lawyers. this assessment is also correct for EU law. the
burden of legal search may be even higher due to insuficiently known EU
terminology that differs from familiar national terminology.
Information retrieval was, and is still, the most obvious and frequently
used method to tackle this research task. Given the limitations of information
retrieval, enhancements are required and thus are the subject of research. Besides
ongoing efforts to improve information retrieval (e.g. SIGIR conferences
( 1 )), legal ontologies as a formalisation tool of knowledge have received
considerable attention in recent years. Ontologies may provide a solution for
this problem because users can reuse formalised knowledge on the text corpus
and its terminology.
In this contribution, an idea for a new approach for structuring legal databases
using ontologies is described and developed based on the example of the database
on European law, EUR-Lex ( 2 ) formerly CELEX (for an overview of EUR-Lex/
CELEX see: 25 years of European law online, Publications Ofice, Luxembourg, 2006).
for the sake of better understanding, let us start with a sketchy deinition.
An ontology is a data model that represents a set of concepts within a domain
ERICH SCHWEIGHOFER
Associate Professor
at Vienna University
(on leave); case handler
for State aid in the
Directorate-General
for Agriculture and Rural
Development, European
Commission, Brussels
( 1 ) website: http://www.sigir.org.
( 2 ) website: http://eur-lex.europa.eu.
136 | 137
01_2007_5222_txt_ML.indd 137 6-12-2007 15:14:02

of knowledge and the relationships between those concepts. A lexical legal
ontology consists of an enlarged thesaurus using knowledge representation
techniques of informatics. for a legal database as a textual representation, an
ontology means extension with semantic data allowing automated reuse of
this information.
REASONS fOR A NEw APPROACh USING ONtOLOGIES
In 1971, the customers of CELEX were information oficers, irst those
within European institutions and later also those of external clients. the technology
was cumbersome but a workable strategy existed. Information oficers
with a high knowledge and experience of structured data (in those days in the
form of card indices) were combined with new technology: an electronic database
and new search options designed according to ‘old’ search options. thus,
a new form of searching was created, that, together with the excitement of
going online, was considered a form of magic.
As information oficers were the clients, only information hints were expected,
not the presentation of all relevant information as documents in full
text. Users took it for granted that results had to be checked and not all information
hints would be really useful.
In 2006, the situation has changed completely. the customers are now
professionals or citizens. technology has greatly improved although the main
concept has remained unchanged. typical users do not have suficient experience
with structured data. Internet search engines such as Google are setting
the new standard. with a keyword search, suficiently relevant documents
should be presented irst following a good ranking algorithm. Users are now
lawyers and citizens that are able to ‘google’ but do not know the particularities
of legal research suficiently well due to the different text corpus and search
environment.
Nowadays, users prefer ‘ready-to-use’ results with high accuracy and
quality and would like to avoid time-consuming and dificult interpretations of
search results. Additional features have to be user-oriented and the output
properly represented. moving into these ‘intelligent’ features constitutes the
main challenge for EUR-Lex for the coming years.
A comparison with the more developed database on legislative procedures
called PreLex ( 3 ) may be illustrative. In comparison with EUR-Lex, PreLex pro-
( 3 ) website: http://www.prelex.europa.eu.
01_2007_5222_txt_ML.indd 138 6-12-2007 15:14:02

WORKSHOP
duces a ‘magic result’. the whole legislative history of a regulation or a directive
is represented in full and presented in a visually attractive way with hyperlinks
to all relevant documents. this ‘intelligent’ output is the result of additional
manual work but not of automatic restructuring of data.
the challenge is therefore to achieve a similar result with an enlarged
EUR-Lex database. Adding metadata to this huge text corpus and introducing
‘intelligent’ interpretation of these documents and data should add ‘magic results’
like those of PreLex.
Legal ontologies can provide the knowledge required to structure legal
knowledge properly. the existing CELEX/EUR-Lex data structures could serve
as a starting point for improvement. Due to space constraints, the question of
improved visualisation is not dealt with. It should be noted that without good
visualisation, the advantages of ontologies would remain invisible and thus
unused.
CELEX/EUR-LEX DAtA StRUCtURES
One should remember the starting point of knowledge structuring in EU
law, the famous CELEX/EUR-Lex data structure (Schweighofer, 1995/2000). ten
main indices and 80 ields had to be addressed in the dificult search language
mistral (e.g. :txt austria+:title &:crossref 157030 &:sub environment). Nowadays,
the commands are replaced by a user-friendly search menu called EUR-Lex
simple search. Another pillar was the CELEX document number, e.g. 1[19]57e030
for Article 30 of the EEC treaty. the CELEX number is still used in EUR-Lex for
document identiication number and cross references between documents.
moreover, documents can be direct linked today via hyperlink, e.g. http://eurlex.europa.eu/LexUriServ/LexUriServ.do?uri=CELEX:11957E030:EN:htmL
( 4 ).
the CELEX structure is focused on documentation. the ield structure
insuficiently relects user needs for a result-oriented output of EUR-Lex. thus,
the structure must be gradually enlarged in order to allow ‘intelligent’ results
from the EUR-Lex database using legal ontologies.
( 4 ) for more information see the EUR-Lex guidelines for linking documents
(http://eur-lex.europa.eu/en/tools/help_syntax.htm).
138 | 139
01_2007_5222_txt_ML.indd 139 6-12-2007 15:14:02

LEGAL ONtOLOGIES
In informatics, ontologies (Gruber, 1993) constitute an explicit formal
speciication of a common conceptualisation with term hierarchies, relations
and attributes that make it possible to reuse this knowledge for automated applications.
the formalisation must be on the one hand suficiently powerful
with regard to knowledge representation, and on the other hand must offer
functionalities for automation as well as tools to be produced automatically
(for lexically based ontologies, see hirst, 2004). thus, a legal ontology is a data
model of the legal order.
Ontologies in law have some particularities due to the legal domain and its
language. this legal text corpus is not inherently structured and a formal taxonomy
does not exist. Legal structuring as such is done by lawyers, in their
minds, and presented and made explicit in their argumentations and writings.
As a product of this process, a legal commentary is considered as the highest
level of this endeavour. the understanding of logic remains also quite different
from the formal logic of computer science: its open legal concepts, inherent
dynamics of law, system models and syntactic ambiguities provide strong impediments
to formalisation.
the motivation for the creation of legal ontologies is evident: common use
of knowledge, examination of a knowledge base, knowledge acquisition, representation
and reuse of knowledge up to the needs of software engineering
(Bench-Capon and Visser, 1997).
wAyS tO LEGAL ONtOLOGIES
As legal concepts constitute the basis for legal ontologies, all endeavours
for semantic representations or the development of thesauri can be seen as
preliminary works. the Semantic web, word nets and conceptual indexing
provide some insights into the proper strategy to achieve a suficiently developed
legal ontology.
As today’s web is mostly text-based, additional semantic data might
bring methods that could be used for the similar task in law. the Semantic
web can be considered as an extension to the current web in providing a
common framework that allows data to be shared and reused ( 5 ). According
to tim Berners-Lee, the Semantic web is ‘not a separate web but an exten-
( 5 ) website: http://www.w3.org/2001/sw/.
01_2007_5222_txt_ML.indd 140 6-12-2007 15:14:02

WORKSHOP
sion of the current one, in which information is given well-deined meaning,
better enabling computers and people to work in cooperation’ (Berners-Lee
et al., 2001).
Information available on the web is semantically tagged and linked using
the technologies of Resource Description framework (RDf), XmL and URIs.
this layer model (Koivunen and miller, 2002) is based on XmL (schema, name
spaces) that offers a structuring of documents and data at the syntactic level.
the next level forms RDf (schema) using the syntax of XmL and providing
clear rules for the production of metadata. RDf describes resources by attributes.
the RDf attributes are deined as a valid vocabulary by the RDf
schema, forming also classes and class hierarchies. the next layer may be a
logical one, an inference machine. In 2004, the w3C published, besides RDf,
the web Ontology Language (OwL) for the development of sets of terms
called ontologies that can be used for supporting advanced web search, software
agents and knowledge management. Besides establishing the framework,
the web has so far not been changed to a semantic representation and offered
a broad high-level structuring of knowledge. In law, the Semantic web constitutes
a tool for representation of domain knowledge but has so far also not
been implemented.
Computer-useable dictionaries may provide information on the world as
such. facts could be checked much more easily. therefore, these projects
would deliver a world description that could be used for identifying factual
situations in law with legal consequences. wordNet is an online lexical reference
system that is an initiative of the linguist George miller. It has been developed
and is being maintained by the Cognitive Science Laboratory at Princeton
University (miller et al., 1990) ( 6 ). Its design is inspired by current psycholinguistic
theories of human lexical memory.
It encodes conceptual relationships between terms by arranging them in a
hierarchical structure. words (nouns, verbs, adjectives and adverbs) and their
short deinitions are grouped into synonym sets (synsets), each representing a
speciic lexical concept. the synsets are linked by a set of different semantic
relations (mainly synonymy/antonymy, hyponymy/hyperonymy, meronymy
and morphological relations to reduce word forms). wordNet aims at supporting
automatic text analysis and AI applications and at providing an intuitively
useable enhanced dictionary. the database of the current version 2.0 contains
( 6 ) website: http://wordnet.princeton.edu/.
140 | 141
01_2007_5222_txt_ML.indd 141 6-12-2007 15:14:02

about 150 000 words organised in 115 000 synsets for 200 000 word-sense
pairs. the wordNet technology primarily aims at linguistic support. As concepts
are deined with natural language terms, no semantic deinitions exist in
a formal language. the deinitions remain vague from a legal point of view.
the motivation for the EurowordNet (EwN) ( 7 ) was the support of monoand
cross-lingual information retrieval. Based on the Princeton wordNet technology,
lexica for eight European languages were developed and connected by
an interlingual index (ILI) (Vossen, 1993). within the EwN, the structure of the
wordNet was supplemented with additional semantic–lexical relations and
three top-level categories. the top level offers 63 semantic distinctions grouped
into three types of entities. they can be accessed by the ILI and form together
the common semantic framework for all European languages. the work on
EwN was inished in 1999 but its framework has been continued by the Global
wordNet Association, which builds on the results of Princeton wordNet
and EwN and provides a worldwide platform for discussing, sharing and interconnecting
wordNets. A standard conversion of the Princeton wordNet to
RDf/OwL has been developed under the auspices of the w3C (wC3, 2006).
the main task of the EU-funded eContent LOIS project (Lexical Ontologies
for legal Information Sharing) was the building of a multilingual legal
wordNet for the purpose of facilitating legal information retrieval. this approach
faced the problem of lack of knowledge of a certain language that prevented
users from formulating queries, and thus from inding relevant results,
but also provided some support to lawyers having to cope with the EU’s linguistic
challenge of 23 oficial languages. Using this framework assured compatibility
of the LOIS wordNets with EwN, and allowed them to function as
an extension of EwN for the legal domain. ten partners from six European
countries (seven universities/research centres and three enterprises) participated
in this project. within the approved project duration of 24 months — 2004
to 2006, around 5 000 synsets were localised for each language involved. the
LOIS project primarily aimed at providing easy access to European legal databases
for legal experts as well as for lay people. further research was envisaged
on improved techniques for information retrieval, on providing document
standards (common XmL standard for the representation of legal documents),
on the commercial use of public-sector information, on showcase applications
( 7 ) the documentation is available at http://www.illc.uva.nl/EurowordNet/docs.html.
01_2007_5222_txt_ML.indd 142 6-12-2007 15:14:03

WORKSHOP
for test and demonstration purposes, and on product placement for integration
of the result into commercial applications (Dini et al., 2005).
Automatic text analysis and conceptual indexing provides support in two
areas: inding concepts based on corpora-based analysis or linking text corpora
to knowledge bases (see for example the following projects: Konterm (Schweighofer,
1999, Schweighofer et al., 2002), Salomon (moens et al., 1997),
flexicon (Smith et al., 1995), SmILE (Brüninghaus and Ashley, 2001), SUm
(hachey and Grover, 2004) or Support Vector machines (Gonçalves and
quaresma, 2005)). Existing techniques — tfxIDf vector document representation
and feature extraction — have proven their feasibility and potential to
structure, classify and describe (label) huge amounts of legal text corpora, in
particular if intellectual improvement is done. however, scaling-up — for example,
deeper investigation of a larger text corpus and use of a more advanced
ontology — has still to be done.
On the one hand, ontologies may offer a solution for structuring of extracted
information. Legal databases could be transformed into a semantic representation
by semi-automatic means, into logical sentences or as process diagrams,
conceptual structures or relationships (e.g. AustLII ( 8 ), CiteSeer ( 9 ). this
semantic representation would not be perfect but would provide signiicantly
easier access to the legal database.
On the other hand, semi-automatic lexical analysis provides invaluable support
for word sense disambiguation or deinitions. In the LOIS project, this tool
was used for extracting multilingual deinitions from the EUR-Lex database.
LEGAL ONtOLOGIES
After important preliminary work (e.g. mcCarty, 1989; hafner, 1977; or
Stamper, 1991), the frame-based ontology fBO of van Kralingen (1995) and
Visser (1995) and the functional ontology fOLaw of Valente (1995) achieved
some prominence. Both were formalised with the description language Ontolingua
(Gruber, 1993) and represent a rather epistemic approach.
the fBO was conceived as a general and reusable legal ontology, which
offered three classes of model primitives, whereby for each unit a frame structure
with all relevant attributes was deined. the types of frames were norm,
action and concept.
( 8 ) website: http://www.austlii.edu.au/.
( 9 ) website: http://citeseer.ist.psu.edu/cs.
142 | 143
01_2007_5222_txt_ML.indd 143 6-12-2007 15:14:03

the aim of fOLaw was the organisation and interconnection of legal
knowledge, in particular with regard to conceptual information retrieval. It
contained six basic categories of legal knowledge: normative knowledge, metalegal
knowledge, world knowledge, responsibility knowledge, reactive knowledge
and creative knowledge. fOLaw was used in follow-up projects. the
central dificulty of fOLaw proved to be the modelling of ‘world knowledge’.
the knowledge gained from fOLaw was used in the e-court project and in the
development of a core legal ontology, LRI-Core (Breuker and hœkstra, 2004).
within the framework of this project, a lexible, multilingual information retrieval
system using heterogeneous sources (audio, video, text) was developed
in the ield of criminal procedure.
Recent projects on legal ontologies are focused on vocabulary, thesauri
and lexical ontologies (Casanovas et al. 2007, LOAIt workshop).
POtENtIAL Of LEGAL ONtOLOGIES
for legal subsumption, factual conditions of an applicable rule have to be
clariied and checked. Legal ontologies should consist of a full description of
the world that would also constitute a statement of all factual situations for a
possible applicability of a norm. this knowledge representation would contain
objects and events from the world, its attributes and relations, and deinitions
and classes.
A representation of the legal order comprises all concepts of the legal system
providing a conceptual description and structuring of the legal order. Applicability
of a rule could be indicated by directly linking a norm to factual
situations described in the world ontology. with such an ontology, an electronic
commentary could be established containing a full description of a norm
with extensive links to related documents. further, it would provide the basis
for automatic legal reasoning.
In artiicial intelligence and law, the main problem has always been the
scaling-up, e.g. the move from a small to a reasonably big application. Legal
ontologies are no exception to this fact. the state of the art has to be described
as still experimental (Casanovas et al., 2007). therefore, a full ontological representation
of EUR-Lex cannot be advised as the theory has not suficiently
developed. however, interesting potential exists for ontological representations
in small and medium-sized sub-documentations of EUR-Lex. Such
smaller applications could be much more easily implemented and might show
the way to an ontological representation of the legal order. further, the present
01_2007_5222_txt_ML.indd 144 6-12-2007 15:14:03

WORKSHOP
focus on text representation and Boolean searching could be maintained but
going far beyond existing uses of legal ontologies like cross-lingual information
retrieval or input reinement.
the main difference to existing features of EUR-Lex would be an automated
reuse of an advanced knowledge representation (e.g. extended EUR-Lex
ield structure). Semantic data would be interpreted and then presented in a
user-friendly output. Support would be provided for the major legal task of
interpretation and legal reasoning.
DEVELOPmENt OPtIONS fOR ONtOLOGIES IN EU LAw
A general transformation of EUR-Lex data into an ontological knowledge
representation could not be recommended as the theory of legal ontologies is
not yet suficiently developed and the scaling-up problem seems to be too dificult.
thus, small full-scale projects should be envisaged. Such projects should
be practical, take into account user needs and require limited additional data
acquisition. further, automated information representation and interpretation
should be feasible at a high level. the focus should be on solving problems in
live situations for professionals and lay users.
At present, navigators through the wealth of information seem to be the
best solution. As such, the ideas behind the following navigators should be
presented in more detail: legislative, citations, layers of the legal order and
terminology.
LEGISLAtIVE NAVIGAtOR
As mentioned, PreLex and its counterpart, Œil, the Legislative Observatory
maintained by the European Parliament ( 10 ), are ine information tools of
the legislative process. with some metadata and intelligent output, EUR-Lex
may also achieve a similar quality to these two specialised databases.
At present, a search for a particular Commission proposal shows the status
of the proposal and — in combination with a cross-reference search — related
documents, e.g. resolutions of the European Parliament, opinions of the European
Economic and Social Committee. Documents not covered by EUR-Lex
but by PreLex are Commission staff working documents, Council working
documents, press releases of the Commission, the Council, the European
( 10 ) website: http://www.europarl.europa.eu/oeil/.
144 | 145
01_2007_5222_txt_ML.indd 145 6-12-2007 15:14:03

Economic and Social Committee, reference notes in the Bulletin of the European
Union and reports of committees of the European Parliament.
the presentation in EUR-Lex remains dificult to understand; it is based
on references and a result list of documents in chronological order. Lay
people or less experienced lawyers often have problems in getting all relevant
information.
In order to overcome these deiciencies and offer the same functionalities
as PreLex, the document corpus must be enlarged with all relevant documents
and additional metadata have to be added. these metadata comprise information
on the status and time of validity of documents in the legislative process.
thus, a presentation with timeline, description of legislative procedure and
additional information on the procedural steps could be created automatically
by rules determining proper data interpretation. the PreLex output representation
could serve as an example of appropriate visualisation. the function of
EUR-Lex would not be only giving hints to useful documents but representing
in real time the status of the legislative procedure of a particular Commission
proposal.
It does not involve much work to improve EUR-Lex with such a feature.
the additional documents are available in electronic format and have only to
be uploaded. (Semi)automatic text analysis may provide the additional metadata.
the presentation of results can follow the PreLex example.
CItAtIONS NAVIGAtOR
Legal citation represents the cross-linked structure of legal acts (Schweighofer,
1999). Legal basis, preparatory acts, amendments, corrections, proposals
for modiications, repeals, related case-law and literature are represented
in EUR-Lex using the CELEX number as the uniform document identiier. the
rich citation structure that is a legacy of CELEX is now somewhat hidden and
is not properly used for searching but more for representing links to related
documents. the simple search of EUR-Lex has still not the functionality of the
old mistral search as the various ields are not suficiently described and combinations
of ields cannot be searched ( 11 ).
( 11 ) List of ields: Amendment_to, CONSLEG, Bf, Instruments_cited, Instruments_cited_in_
case_law, case_affecting, Earlier_related_instruments, Affected_by_case, Legal_basis,
Amended_by, Subsequent_related_instruments. the online guide does not contain a
description of these ields.
01_2007_5222_txt_ML.indd 146 6-12-2007 15:14:03

WORKSHOP
As a irst step, the existing representation of relations should be restructured
using URI style references. (Semi)automatic text extraction tools may
help in inding references in the various documents. the structure of citations
should then be improved, focusing on the main types of legal basis, amendments
and repeals, preparatory acts and proposed modiications, case-law and
literature. Relations should be represented with a graph, with the document in
question in the centre and with arrows linking to boxes with the various types
of citations. An eficient ranking algorithm should select the most important
citations.
while the additional metadata does not constitute a big problem and a
more appropriate visualisation may be introduced quite easily, the ranking algorithm
remains the biggest challenge. So far, the otherwise helpful weighting
algorithms have not worked properly in a legal environment. however, a better
structured visualisation of citations provides suficient value to the user in
order to focus on this improvement and postpone the weighting algorithm for
later implementation.
LAyERS Of thE LEGAL ORDER NAVIGAtOR
A weak point of EUR-Lex is the missing (semi)automatic consolidation of
legal acts. main documents, their amendments and corrections have to be consolidated
by the user. Luckily, more and more acts are also available in a consolidated
version. this work has greatly improved the availability of workable
text versions; however, it requires a lot of resources.
A restructuring of the text corpus would provide the basis for a
(semi)automatic consolidation. the Austrian legal database RIS (Rechtsinformationssystem
des Bundes) ( 12 ) could serve as a role model. In this database, the
various laws are segmented into the subelements of sections, articles or annexes.
the time of validity is stored per document segment. Standard searches
also retrieve documents with a given date of validity, e.g. as a default value the
day of searching.
European documents are not as formally structured as Austrian laws but
small adaptations will solve this problem. Recitals of the preamble could be
treated like articles. A basic document would contain the title, legislative history
and additional information like citations of the bundle of documents of a
regulation or directive.
( 12 ) website: http://www.ris.bka.gv.at.
146 | 147
01_2007_5222_txt_ML.indd 147 6-12-2007 15:14:03

As EU law is more complex, different types of applicability have to be
developed, in particular on the validity and the direct applicability of the act.
As the work for restructuring such a huge database as EUR-Lex is quite high, a
(semi)automatic solution has to be developed in order to pursue this approach.
Additional research is necessary for deining the various layers of validity and
applicability.
tERmINOLOGy
At present, four instruments of terminology are available in EUR-Lex: the
Eurovoc thesaurus, classiication headings, subject matters and case-law directory
code. the best legal quality can be found in the case-law directory code.
quite useful classes of legislation are available with the classiication headings.
the Eurovoc thesaurus has its usefulness in general searching. however, weaknesses
are evident in legal searches due to an insuficient number of legal concepts.
thus, Eurovoc should be extended with a lexical ontology that contains
the required high number of legal terms. A conceptual structure must also be
developed for these terms. Such an instrument would greatly improve linguistic
support for searching. Deinitions of and relations between the concepts
would help lay users and less-experienced lawyers. Such a terminology could
be used for a reinement of queries, cross-linguistic retrieval and a multipurpose
dictionary.
So far, lexical ontologies on European law have been limited to speciic
purposes (e.g. the LOIS project). A full-scale lexical ontology would require a
high level of resources that need a stronger focus of lawyers on terminology
questions. thus, hopes rest on (semi)automatic analysis in order to extract the
suficient number of concepts from a representative text corpus.
CONCLUSIONS
today, EUR-Lex is an excellent text provider but an insuficient metacontent
provider. the main reason is that legal ontologies are insuficiently
developed for such a large application. It is not advisable to move immediately
to a full-scale legal ontology. Intermediate steps are very helpful and much
easier to take. Examples are the described navigators: legislative/juridical, citations,
layers of the legal order and terminology.
01_2007_5222_txt_ML.indd 148 6-12-2007 15:14:03

WORKSHOP
REfERENCES
Bench-Capon, t. J. m. and Visser, P. R. S., ‘Ontologies in legal information systems: the
need for explicit speciications of domain conceptualisations’, Proceedings of the
6th ICAIL (melbourne, Victoria, AU, 1997), ACm Press, New york, Ny, 1997, pp.
132–141.
Berners-Lee, t. et al., ‘the semantic web’, Scientific American, Vol. 284 (05/2001), Scientiic
American Inc., New york, Ny, 2001, pp. 34–43.
Breuker, J. and hoekstra, R., ‘Direct: ontology-based discovery of responsibility and
causality in legal case descriptions’, Proceedings of the 17th JURIX (Berlin, DE,
2004), IOS Press, Amsterdam et al., NL, 2004.
Brüninghaus, S. and Ashley, K. D., ‘Improving the representation of legal case texts with
information extraction methods’, Proceedings of the 8th ICAIL (St. Louis, mO,
2001), ACm Press, New york, Ny, 2001, pp. 42–51.
Casanovas, P., Biasiotti, m. A., francesconi, E. and Sagri, m. t. (eds), Proceedings of
LOAIt 07, II. workshop on legal ontologies and artiicial intelligence techniques,
ICAIL-07 workshop, Stanford, CA, 2007.
Dini et al., ‘Cross-lingual information retrieval using a wordNet architecture’, Proceedings
of the 10th ICAIL (Bologna, It, 2005), ACm Press, New york, Ny, 2005, pp.
163–167.
Gonçalves, t. and quaresma, P., ‘Is linguistic information relevant for the classiication
of legal texts?’, Proceedings of the 10th ICAIL (Bologna, It, 2005), ACm Press,
New york, Ny, 2005, pp. 168–176.
Gruber, t. R., ‘A translation approach to portable ontology speciications’, Knowledge
Acquisition, Vol. 5/2 (1993), Academic Press, London et al., 1993, pp. 199–220.
hachey, B. and Grover, C., ‘A rhetorical status classiier for legal text summarisation’,
Proceedings of the ACL-04 text Summarization Branches Out workshop (Barcelona,
ES, 2004), pp. 35–42.
hafner, C. D., An information retrieval system based on a computer model of legal knowledge,
UNI Research Press, Ann Arbor, 1977.
hirst, G., ‘Ontology and the lexicon’, in S. Staab and R. Studer (eds), Handbook on ontologies,
Springer, Berlin-heidelberg, DE, 2004, pp. 210–229.
Koivunen, m.-R. and miller, E., ‘w3C semantic web activity’, Proceedings of the Semantic
web Kick-off Seminar (helsinki, fI, 2001), hIIt Publications 2002/1, helsinki,
fI, 2002, pp. 27–-43, available free at http://www.w3.org/2001/12/semwebin/w3csw.
Kralingen, R. w. van, ‘frame-based conceptual models of statute law’, Ph.D. thesis,
University of Leiden, the hague, NL, 1995.
148 | 149
01_2007_5222_txt_ML.indd 149 6-12-2007 15:14:03

mcCarty, L. t., ‘A language for legal discourse: I. Basic features’, Proceedings of the
2nd ICAIL (Vancouver, BC, Canada, 1989), ACm Press, New york, Ny, 1989,
pp. 180–189.
miller, G. A. et al., five papers on wordNet, CSL Report 43, Princeton University 1990:
Cognitive Science Laboratory, ftp://ftp.cogsci.princeton.edu/pub/wordnet/5papers.ps.
moens, m.-f. et al., ‘Abstracting of legal cases: the Salomon experience’, Proceedings of
the 6th ICAIL (melbourne, Victoria, Australia, 1997), ACm Press, New york, Ny,
1997, pp. 114–122.
Ofice for Oficial Publications of the European Communities, 25 years of European law
online, Luxembourg 2006.
Schweighofer, E. et al., ‘Improvement of vector representation of legal documents with
legal ontologies’, Proceedings of the 5th BIS (Poznan, PL, 2002), Poznan University
of Economics Press, Poznan, PL, 2002.
Schweighofer, E., wissenrepräsentation in Information Systemen am Beispiel des EU-
Rechts, Dissertation, Universität wien (published wUV; with Addendum XmL),
1995/2000.
Schweighofer, E., Legal knowledge representation: automatic text analysis in public international
and European law, Kluwer Law International, the hague, 1999.
Smith, J. C. et al., ‘Artiicial intelligence and legal discourse: the flexlaw legal text management
system’, Artificial Intelligence and Law, Vol. 3/1-2 (1995), Kluwer, Dordrecht
et al., NL, 1995, pp. 55–95.
Stranieri, A. and Zeleznikow, J., Knowledge discovery from legal databases, Dordrecht,
Springer, 2005.
Valente, A., Legal knowledge engineering: a modelling approach, IOS Press, Amsterdam, NL,
1995.
Visser, P. R. S. ‘Knowledge speciication for multiple legal tasks: a case study of the interaction
problem in the legal domain’, Computer Law Series, Vol. 17, Kluwer Law
International, the hague, NL, 1995.
Vossen, P. (ed.) EurowordNet General Document (LE2-4003, LE4-8328), inal document
(Version 3), 1993, available free at www.illc.uva.nl/EurowordNet/docs.html.
w3C (mark van Assem, m., Gangemi, A. and Schreiber, G. (eds.)) RDf/OwL Representation
of wordNet, 2006, http://www.w3.org/tR/2006/wD-wordnet-rdf-20060619/.
01_2007_5222_txt_ML.indd 150 6-12-2007 15:14:03

TEXT MINING
1. INtRODUCtION
the growth of electronically available information on the Internet has
been enormous and is still increasing. In february 2006, the Google search
engine offered access to more than 25 billion web pages. Eight years before,
the igures were relatively low, with about 320 million pages.
In its introduction, the EUR-Lex website currently mentions about 1.8 million
documents in various languages which are accessible via this portal. In
2005, which was supposed to be a quiet year, more than 2 100 regulations
were adopted in European legislation which meant over 44 000 new documents
in EUR-Lex. Although these amounts seem relatively small in comparison
with the igures on the Internet, a human client of the EUR-Lex system will
be lost without an eficient support from corresponding electronic information
retrieval systems.
Nevertheless, it has to be underlined that, since computers irst started to
store and retrieve information, individuals have been confronted with more
and more information items. thanks to information technology research, retrieval
methodologies have become more and more sophisticated and offered
easier and better possibilities for information retrieval.
It is today certainly understandable that users of information retrieval systems
are very demanding about the quality of the results, and no longer by the
quantity. what is the advantage when a search engine on the web confronts a
user with millions of hits? Sometimes it helps if the retrieval parameters are
reformulated; Boolean operators may also help to exclude some websites or at
least to determine which kind of answer its a given request better. In many
cases, however, this will not necessarily reduce the number of answers, but
will perhaps place the most relevant answers on top of the list. however, this
is not really an improvement of the retrieval result.
Another problem in the same context is that accessible systems or portals
are so complicated that users do not necessarily have the knowledge and experience
to make the engines extract relevant data. this can be supported by the
HOLGER BAGOLA
Head of the Formats
and Linguistic Tools Section,
Publications Office
150 | 151
01_2007_5222_txt_ML.indd 151 6-12-2007 15:14:05

development of user interfaces, which by means of taxonomies lead the user
to the searched domain of interest (Dörre, Gerstl and Seiffert, 2004, p. 480).
One of the biggest challenges of information retrieval systems is the fact
that most (if not all) of the documents — in the broader sense of the word —
are written in a natural language ( 1 ). Among others these communication systems
are characterised by a certain number of different possibilities to refer to
the same extralinguistic facts. Sometimes it is even a good method in a communication
situation to reformulate a statement in order to show the understanding
of the original one and/or to get a conirmation of it. In other cases,
certain facts are paraphrased without a speciic keyword being used. for example,
if an author describes a nice lady in white clothes sitting on a glass
bowl, it could mean that he is talking about luck. this example of a medieval
allegory, which today is probably only understood by some experts, proves
another characteristic of human language: it changes over the course of time.
New expressions appear and others disappear or change their meaning.
It has also to be kept in mind that the items of our vocabulary are not only
in a certain relation with the concept of an extralinguistic fact, but are related
to each other as well. this is why, for instance, in good dictionaries you ind
hints about synonyms or antonyms to a given expression. But in the context
of environment protection, for example, it could also be of interest to retrieve
information on air pollution. this case indicates the complexity of the relations
between expressions, and it may be doubted if such word ields are ever
complete.
Such circumstances obviously make an automatic retrieval of performing
information extremely complex. So documents need to be analysed by means
of scientiic linguistic methodology which goes far beyond the still widespread
approach of simple text indexing and clearing. the still ‘young’ domain of text
mining tries to develop appropriate methods to support the digging for information
within natural language documents.
text mining is sometimes also referred to as ‘text data mining’ or ‘knowledge
discovery in text’. In general it deines the process of retrieving information
in texts. this is the most important difference between data mining and
text mining. while data mining procedures try to extract relevant information
from structured databases, text mining concentrates on unstructured text doc-
( 1 ) Some attempts to translate documents into a more formal language — Interlingua —
were not very successful. See the comments by hutchins (1986, Chapter 10).
01_2007_5222_txt_ML.indd 152 6-12-2007 15:14:05

WORKSHOP
uments. the tasks and objectives for the analysis process are more or less the
same (Dörre, Gerstl and Seiffert, 2004, p. 480).
Information is typically identiied through processes discovering patterns
and relations mainly by means of statistical pattern learning. texts are generally
regarded as unstructured data in contrast to database information, which
is supposed to be structured. text mining usually involves the process of structuring
the input text by ‘parsing’, which is completed by the addition and/or
removal of linguistic features. this restructuring of data permits the derivation
patterns as well as evaluation and interpretation of the output. the quality of
text mining is usually judged on the combination of relevance, novelty and
tractability. typical text-mining tasks include text classiication, text clustering,
concept or entity extraction, document summarisation and modelling of entity
relations.
text-mining processes may be described as a subsequent low of activities.
By means of statistical algorithms, the key terms of a textual entity are
identiied. Comparison with entries in ontologies offers possibilities to group
those texts together with similar ones. In this way, a basis of knowledge is created
and extended after analysing other documents.
An example will show the complexity of the necessary methods. Imagine
that a document contains the German word Birne ‘pear’. It has to be taken into
account that the use of this term could be an ellipsis or a metaphor. that leads
us to the following virtual classes, which distinguish from each other by the
different meanings of the key term:
(1) a kind of fruit,
(2) the tree which produces the fruits (‘pear tree’); this is the elliptic use for
Birnenbaum,
(3) the wood of a pear tree which is used for the construction of furniture; this
is an ellipsis for Birnenholz,
(4) an electric bulb which in many cases has a form resembling a pear; this is
a metaphor well established in the German vocabulary and at the same
time an ellipsis for Glühbirne,
(5) ironically the head of a human being which in certain stylistic contexts
may be compared with a pear; in that case it could be regarded as a metaphor.
Although the last one of these variants only has to be taken into account
depending on the stylistic context, the other ones need deeper analysis so that
the documents concerned can be related to similar ones. In the irst case, this
could consist of references to other types of fruit or foods. If the document
152 | 153
01_2007_5222_txt_ML.indd 153 6-12-2007 15:14:05

deals with fruit plantations, it should be classiied in the second group. the
third group is relevant if, for instance, materials for furniture are discussed. the
fourth group is based on a metaphoric application of the term; it deals with
electricity or artiicial illumination.
2. mEthODOLOGICAL PARtICULARItIES Of tEXt mINING
the methodologies of text mining include technologies developed in the
context of computer linguistics or linguistic informatics. Actually it is underlined
that text mining has led to a revival of the corresponding ideas and features.
In particular, mathematical and statistical approaches are supposed to be
of high importance. the identiication of so-called stop words — functional
words which describe relations between terms without having a special meaning
of their own — and the calculation of word frequencies are of basic interest.
they contribute to the analysing of those patterns which are essential to
the relevant meaning of the text. the methods may be paraphrased, as done
by hippner and Rentzmann (2006):
(text) ‘mining methoden: Nachdem terme aus den textdokumenten extrahiert
worden sind und die textuellen Daten somit eine Struktur erhalten
haben, können Verfahren angewandt werden, die aus dem klassischen
Data mining bekannt sind: texte können automatisch vorgegebenen Kategorien
zugeordnet werden (Klassiikation) oder sie können so gruppiert
werden, dass ähnliche texte zusammengeführt werden (Segmentierung).
Ebenso kann das gemeinsame Auftreten von termen analysiert werden
(Abhängigkeitsanalyse).’ ( 2 )
In an extension of the abovementioned example, this could lead to the
recognition of ‘pear’ (Birne) being a central term of the document. It is, however,
still not known which of the different meanings is concerned. Ontologies
could help to take other terms into account as well, and to inally arrive at a
clear document classiication. the network of an ontological description of the
concepts around Birne is certainly much more complex than this illustration
might express:
( 2 ) translation: ‘text-mining methods: After the extraction of expressions from text
documents thus giving textual data a structure, methods can be applied which are wellknown
from classical data-mining: texts can be associated to predeined categories
(classiication) or they can be grouped in a way that similar texts are brought together
(segmentation). It is also possible to analyse the existence of expressions in common
(analysis of dependencies).’
01_2007_5222_txt_ML.indd 154 6-12-2007 15:14:05

WORKSHOP
In order to create more eficient ontologies, the granularity has to be much
iner. thus the distinction between wood that is used for construction and
wood which normally is not used in this context could be clearer, as the
number of direct links could be reduced. In the abovementioned example this
distinction is made by means of multiple links from the central term: ‘pear’,
‘oak’ and ‘beech’ have links to ‘tree’ as well as to ‘wood’ whereas ‘apple’ is
only related to ‘tree’.
It is obvious that one of the main problems of this approach is the availability
of corresponding ontologies. In fact, onomasiological dictionaries which
have been created together with knowledge schemata in the last decades could
be of high interest ( 3 ). Unfortunately the cooperation between the classical
linguistic lexicography and modern computer-assisted and oriented research is
not always very intensive.
the research for linguistic patterns is also supported by the use of thesauri.
the distinction between main terms and synonyms or near-synonyms
( 3 ) See, for example, the onomasiological dictionary project of the Romance languages by
henri Vernay (1991).
154 | 155
01_2007_5222_txt_ML.indd 155 6-12-2007 15:14:05

allows for identifying the concepts of a document. In fact, both approaches
should be used subsequently to arrive at a clear and solid classiication of texts.
the documents actually are instantiation of different concepts.
text-mining technology is not limited to the basic ideas which are expressed
at the surface of a given text, but discovers other ones as well. thus
documents can be related to each other although the key terms used are different.
for example, a document handling the legislation on ‘environment protection’
can be classiied together with documents treating problems of ‘air pollution’.
Linguistic statistics offer basic methods for the distinction of relevant and
less- or non-relevant components. Regularities concerning the frequency of
words and word forms in a given language allow the description of linguistic
relations. this approach combined with a probabilistic linguistic model is one
of the fundaments of text mining.
One of the most important methods analyses the differences of the vocabulary
of a given text with a reference text collection which is supposed to
represent the general vocabulary of a language. After determining the frequencies
of all word forms or tokens in both the text and the reference collection,
there are four classes which help to identify a word:
(1) words which exist in the analysed text, but are not part of the reference
collection: there is a very high probability that these words belong to the
speciic vocabulary of the domain which the text deals with;
(2) words which exist both in the analysed text and the reference collection,
but with the relative frequency in the analysed text being higher than in
the general collection: if a predeined threshold value is exceeded, it is
probable that these words also belong to the speciic terminology of the
domain which the text speaks about;
(3) words which exist both in the analysed text and the reference collection
and whose relative frequencies are more or less the same: in general these
words are necessary for the functioning of a language, they do not regard
speciic subjects;
(4) words which exist in the reference collection with higher relative frequencies
than in the analysed text: with a very high probability these words do
not contain terms of the subject matter treated by the analysed document.
the advantage of this method is that the key terms of a given text can be
identiied with a relatively high probability. the next step could consist of the
classiication of the document with documents which deal with similar sub-
01_2007_5222_txt_ML.indd 156 6-12-2007 15:14:05

WORKSHOP
jects. the problem, however, is the collection of reference texts. As the general
use of a language should be represented, a clear deinition of the borders between
a general and a speciic vocabulary has to be found. But in the daily
world, which is more and more inluenced by new technologies, this deinition
is not evident.
A similar problem exists for the deinition of threshold values. their exactness
has a direct impact on the usefulness of the second one of the abovementioned
word classes. this is of particular importance with respect to the generalisation
of speciic vocabularies in the everyday language, which tends to
minimise the irst class.
to overcome such limits, additional methods have to be implemented.
the probabilistic language model, which is based on syntactical and semantic
analyses, leads to the deinition of rules which limit the number of combinations
of linguistic entities. the example The bone eats a dog, although syntactically
correct, has to be rejected as bone cannot be the acting part in the context
of the verb eat. A similar rule will exclude Birne ‘bulb’ from being the object of
the same verb ( 4 ).
Various other methods will have to be used to reine text-mining analyses.
the cooperation of different scientiic disciplines will be necessary in order to
deine relevant vocabularies.
3. ImPLEmENtAtION Of tEXt mINING
Before text-mining methodologies can eficiently contribute to the acquisition
of knowledge, an enormous amount of preparatory work has to be done.
In particular, dictionaries have to be created which contain suficient information
for the text analysis as well as the necessary interlinking such as described
by ontologies.
this may be one of the reasons why text mining is established for speciic
subject matter. the life science domain is particularly active. It is supposed to have
the largest user community and the fastest-growing literature. the fraunhofer Institut
in Bonn-St Augustin, Germany, organises an annual conference where representatives
from various subjects report on the evolution of their projects.
A project which is at a state of relatively high maturity is Biotem (Deutsches
virtuelles Centrum für text mining in der Biomedizin (the German vir-
( 4 ) Leaving aside some sensational performers who lead their audience to believe that they
are really eating bulbs.
156 | 157
01_2007_5222_txt_ML.indd 157 6-12-2007 15:14:06

tual centre for text mining in biomedicine)). It offers the automatic analysis of
publications of any kind in biology and medicine. furthermore it helps to classify
the electronically available, but mostly unstructured, information on patients.
thus, account can also be taken of relations which until now had not
been taken into account. In the United States of America, the National Library
of medicine has used text mining for 15 years with great success.
the British government has supported the UK National Centre for text
mining with a GBP 1 million grant. the Japanese national parliament recently
decided to establish a centre for text mining in biology. the list can go on, and
shows that the technology is considered to be well advanced. Drew Robb
(2004) gives an impressive list of projects world wide, which handle tremendous
amounts of data.
4. tEXt mINING IN COmPARISON wIth OthER INfORmAtION-
REtRIEVAL mEthODOLOGIES
Information retrieval has not only become a discipline since the widespread
use of computers, particularly personal computers. Some of the methodologies
are in fact even older than computers. Some of these methodologies
will be described, together with an analysis of the sort of advantages that textmining
technologies offer.
Boolean retrieval
Boolean retrieval is widely used, mostly because of its simple syntax.
terms are researched and may be combined with the operators AND (∧),
OR (∨) and NOt (¬). Although there are possibilities to modify the priorities
of combinations which might result in rather complex requests, it is not too
dificult to validate the syntax of a command. however, it is impossible to
control the semantics of a request; for instance, it is not possible to detect
terms which exclude each other. In many applications which are based on this
technology, the syntax is extended by additional operators for comparison
such as > (greater than), < (less than), = (equal), ≥ (greater or equal), ≤ (less or
equal) and ≠ (not equal).
terms may be related to ields of structured data in a database or refer to
words or expressions within unstructured data. the eficiency of Boolean retrieval
is often improved by collecting keywords in so-called inverted lists.
Each expression is accompanied by references to the documents from which it
was extracted. the problem is, however, that these lists take up a lot of storage
01_2007_5222_txt_ML.indd 158 6-12-2007 15:14:06

WORKSHOP
space and their maintenance may be very time-consuming, because in many
cases the inverted indexes have to be regenerated. this is why the expressions
of an inverted list are reduced to those terms which have been collected in a
controlled vocabulary.
Boolean retrieval has some weaknesses. In particular, the quality of a retrieval
result is very dificult to control. furthermore, in some cases the retrieval
is considered to be too strict. A request such as A ∧ B ∧ C delivers results
only if the data contains all three expressions; the result is empty if only
two conditions are fulilled.
Vector model
In general, vector model-based systems lead to the best result in retrieval
processes. Instead of focusing on single-index items from a microscopic point
of view, the vector model starts from a macroscopic perspective. It supposes
that documents are characterised by the statistical distribution of the terms.
when a retrieval is started, a vector model-based system tries to identify those
documents which suit the query best with regard to the statistics of the concerned
terms. to do so, formulae have been developed which help to calculate
the similarity between various documents.
the calculation of distances between text documents is based on high dimensional
vectors of features. the features identiied in all documents of a
given collection are extracted. the sum of all these features creates a feature
space. On the basis of predeined selection criteria, the number of these features
is reduced. the elimination of so-called stop words is one of a number of
applied techniques. Another one determines — by means of statistical analysis
— particularly high or low frequent terms. At the end of this phase, which
is generally called feature reduction, the feature space has n dimensions.
It is now possible to describe each text of the collection by a vector. the
value of an element in each dimension, v j , is deined by the feature matrix
which corresponds to the document. the concrete value of v j depends on the
applied methodology. Sometimes it is suficient to signal the presence or absence
of a feature by the igures 0 and 1. In other cases, the absolute of normalised
frequence is preferred. Normalisation is generally necessary to compensate
for the varying length of documents. the distance between two texts can
now be determined on the basis of the corresponding vectors. Simple values
depend on the distance between the points deined by the vectors in an
n-dimensional space or by the angle which is formed by the vectors. In general,
the values are normalised to results which are placed between 0 and 1, for
158 | 159
01_2007_5222_txt_ML.indd 159 6-12-2007 15:14:06

example by using cos (α) instead of the angle α. An inverted value is often used
to express a low distance; in that case a value of 1 signals the complete identicalness
of the vectors (Dörre, Gerstl and Seiffert, 2004, pp. 491–2).
Indexing enforced by computer linguistic technologies
Simple lists of word forms turned out to be rather ineffective because
terms are only recognised if they correspond exactly to the registered occurrence.
Computer linguistics have developed algorithms which allow text forms
to be reduced to basic forms, normally the nominative singular for nouns and
the ininitive for verbs. this process is generally known as lemmatisation and
groups together all word forms found in a text document under the corresponding
forms which introduce articles in a dictionary.
Although lemmatisation is already an important step in various kinds of
analyses, it is not always satisfying when trying to extract information.
words — or even lemmata — are supposed to represent the same concept if
their roots are the same. the following table gives an example, and distinguishes
between formal roots for which the ending morpheme is suppressed,
text forms, lexical roots or lemmata, and the root of the term.
Text form Lexical root (lemma) Formal root Root
absorb
absorbed
absorbing
absorb
absorbs
absorber
absorbers
absorber
absorbable
absorbably
absorbable
absorbance
absorbances
absorbance
absorbancy
absorbancies
absorbancy
absorbent
absorbents absorbent
absorbently
absorption
absorptions
absorption
absorptive
absorptively
absorptive
Source: ferber, 2003, p. 43.
absorb
absorbab
absorbanc
absorbent
absorption
absorptiv
absorb
01_2007_5222_txt_ML.indd 160 6-12-2007 15:14:06

WORKSHOP
the success of the implementation depends on the language. for English,
for example, there are very eficient algorithms for lemmatisation as well as for
the reduction to roots. Other languages are more complicated, in particular
those which function with discontinued word forms such as German and
Dutch.
Der Mann holt den Koffer am Bahnhof ab ⇒ abholen
In other languages there are not only modiications of the beginning or
end, but there are inlexions as well. In any case, these computer linguistic
technologies are auxiliaries for retrieval approaches, which are based on
Boolean methods or even go much further.
Classiication
this method is supposed to be the oldest one, particularly because it was,
and still is, used in libraries. the text objects are classiied according to the
themes they deal with. most of these systems have a hierarchical structure.
One of the most sophisticated classiication systems is the Universal Decimal
Classiication, which in its core version consists of over 56 000 terms accompanied
by another 13 000 common auxiliaries. the latter cover subjects
such languages, etc. the complete system currently contains about 220 000
terms. the system was created by two Belgian librarians at the end of the 19th
century. An important characteristic of this system is that various categories
may be linked to each other. thus it is possible to treat quite sophisticated research
requests.
this is an extract from the Universal Decimal Classiication.
0 Generalia
00 Prolegomena. fundamentals of knowledge and culture.
Computer science
001 Science and knowledge in general. Organisation of
intellectual work
002 Documentation. Books. writings. Authorship
003 writing systems and scripts. Including: signs and symbols
004 Computer science and technology. Computing
004.2 Computer architecture
004.3 Computer hardware
004.4 Software
004.5 human–computer interaction
>>>
160 | 161
01_2007_5222_txt_ML.indd 161 6-12-2007 15:14:06

004.6 Data
004.7 Computer communication
004.8 Artiicial intelligence
004.9 Application-oriented computer-based techniques
005 management (Revision from 2001)
006 Standardisation of products, operations, weights,
measures and time
007 Activity and organising. Information. Communication
and control theory generally (cybernetics)
008 Civilisation. Culture. Progress
009 humanities. Arts subjects in general
01 Bibliography and bibliographies. Catalogues
02 Librarianship
03 General reference works. Encyclopaedias
050 Serial publications. Periodicals (their function, business
and editorial management)
06 Organisations and other types of cooperation. Including:
Associations. Congresses. Exhibitions. museums
070 Newspapers. the press. Including: Journalism
08 Polygraphies. Collective works
09 manuscripts. Rare and remarkable works
1 Philosophy. Psychology
2 Religion. Theology
3 Social sciences
4 Unassigned
5 Natural sciences
6 Technology
7 The arts
8 Language. Linguistics. Literature
9 Geography. Biography. History
It has to be underlined that the classiication systems are static, although
relations may be indicated. An object can only be stored in one category; for
other aspects, additional methods have to be applied. the classiication
schema has to be ixed before its irst use; later modiications are very dificult
to implement, because there is a risk of having to reclassify data already stored.
Nevertheless, the classiication methodology was revitalised by the Internet, as
it is quite simple to create a hyperlink system on this basis. most search engines
offer thematic categories which simplify the research of the users.
01_2007_5222_txt_ML.indd 162 6-12-2007 15:14:06

WORKSHOP
Thesauri
thesauri are controlled vocabularies with a hierarchical structure. In this
respect, they reuse the same approach as the classiication methodologies, but
on the level of the components of natural speech. the main items which are
registered in a thesaurus are generally called descriptors, as they are used to
describe the contents of a textual object. the entries in the thesaurus are reined
by a sophisticated system of links to broader terms, narrower terms,
synonyms and related terms.
the advantage of the use of thesauri is that research can automatically be:
• redirected from registered synonyms to the corresponding descriptors;
• limited to narrower terms if the number of results exceeds a certain threshold;
• extended to broader and/or related terms if the results are too poor.
Although a big step forwards, thesauri also have their shortcomings.
the quality of the research results depends on the quality of the use of the
descriptors. furthermore, only the main concepts of a text will be taken
into consideration. So there is always the risk that texts are only retrieved
in the context they are prepared for, while other present concepts are not
found.
Semantic networks
the idea of thesauri is driven forward by the technologies in the context
of a semantic network. they are mainly based on the use of ontologies, which
are comparable to thesauri, but instead of linking terms, they describe the relations
of concepts. the example overleaf illustrates this approach in comparison
with thesauri.
the example is simpliied, but it is obvious that the description of the nature
of the relation is much more useful than a pure link.
the evaluation of these descriptions allows for some logical conclusions
which can automatically be made on such basis. In semantic networks, heritage
is an important aspect. So, on a deeper level, the properties of the hyperonymic
level are inherited and are also available for association with other
concepts.
As the work for the creation and maintenance of thesauri is already very
time-consuming and complex, the elaboration of ontologies is even more
complicated. this is why this technology is not yet used for big projects or
large amounts of data, but for small excerpts. the spanning of the Internet by
162 | 163
01_2007_5222_txt_ML.indd 163 6-12-2007 15:14:06

Thesaurus
Descriptor:
pear tree
…
Entry
Descriptor:
pear
Entry
Related term:
pear
Related term:
pear tree
Ontology
Entry
Instance:
pear tree
Instance:
pear
Entry
Produces:
pear
Produces:
wood
Instance:
wood
Entry
Originates from:
pear tree
Originates from:
pear tree
Is used for:
furniture
a semantic web is a very ambitious and interesting idea, but in spite of various
analyses it is and will stay a dream for a long time to come.
Text mining
In fact, text mining is a conglomeration of the different methodologies
which are brought together and which beneit from synergetic effects. But text
mining is even more ambitious. the analysis of unstructured data is supposed
to be automatic, which in fact means independent from any particular preparation
of the text documents. A text-mining system is furthermore supposed to
be extensible, which means that, just by training, new rules can be integrated
in the analysing processes. the most important advantage would be that automatic
procedures can detect relations which would not otherwise have been
highlighted by a human expert when categorising a document.
If the objects to be analysed are not prepared, i.e. indexed by means of
controlled vocabularies or classiied, the text-mining system must have access
to sophisticated resources. It requires formalisms for the representation of
01_2007_5222_txt_ML.indd 164 6-12-2007 15:14:07

WORKSHOP
knowledge which are reusable. this knowledge has to be integrated in the
existing knowledge base. During the retrieval process, relations between identiied
terms have to be validated and evaluated. All these preparations or preliminary
tasks are language dependent. So the knowledge base, at least, has to
be conceived in a way which makes it independent of a given language.
5. tEXt mINING IN LEGISLAtIVE DAtA?
the fourth symposium on text mining organised by the fraunhofer Institute
concentrated on legislative data. Various projects from different countries
were presented and may be considered to be very ambitious. As all of them are
concentrated on limited subjects — that does not exclude that huge amounts
of data are treated — the speakers could talk about their success. however, it
has also to be mentioned that the real use can hardly be veriied.
Applying text mining to a complete legislation, even a rather young one
such as the legislation of the European Union, is therefore extremely ambitious.
the different subjects are characterised by specialised vocabularies and
terminologies. the reusability of tools, dictionaries or rules for the evaluation
of relations is limited. Consequently, the design and development will take up
very large amounts of time and resources.
Nevertheless, legislative texts have a big advantage in comparison with
other document types. Even if their structure is not explicitly marked up, they
have an inherent structure which, in many cases, has a long tradition. Parts of
the document structure can be identiied by evaluating the use of speciic key
terms. As a consequence, an analysis in the context of information retrieval
could be directed to and concentrated on those components. therefore, in
many cases, the application of text-mining technologies to titles of legislative
documents could be suficient to classify acts or associate descriptors from a
thesaurus without the intervention of a human operator.
the abovementioned complexity and possible dificulties should not be
regarded as reasons for not advancing towards automation processes. As textmining
systems are in general very modular, it should be foreseen that smaller
components could be implemented in existing systems. this could help to
validate the eficiency of the modules and, at the same time the operational
systems could be improved. Implicitly, this also means that the quality of the
retrieval results should improve as well.
164 | 165
01_2007_5222_txt_ML.indd 165 6-12-2007 15:14:07

SELECtIVE BIBLIOGRAPhy
Abendroth, Jan. 2003. Entwurf und prototypische Realisierung eines Dokumenten Management
Systems mit RDF-Metadatenverwaltung. Diplomarbeit. technische Universität Ilmenau.
xi + 118 pp. [http://www.imt.tu-ilmenau.de/~schoen/prakinf-studentenarbeiten/
da-2003-abendroth.pdf (last visit: 25.10.2006)]
Baumüller, Andreas. 2002. Topic Maps zur semantischen Strukturierung von Wissensbasen im
Knowledge Management. Diplomarbeit. technische fachhochschule wildau. 94 pp.
[turku.wi-bw.tfh-wildau.de/~cmueller/Diplomarbeiten/Baumueller_Andreas.pdf
(last visit: 25.10.2006)]
Bergmann, tobias. 2005. Regelbasierte Dokumenten-Klassifikation (Verwendung von Text zur
maschinellen Vorhersage) 14 pp. [http://www.informatik.uni-ulm.de/ni/Lehre/SS05/
Proseminartextmining/ausarbeitungen/Bergmann.pdf (last visit: 6.9.2006)]
Bloehdorn, Stephan; Cimiano, Philipp; hotho, Andreas; Staab, Steffen. 2005. ‘An
ontology-based framework for text mining’. In: mehler/wolff. 2005a. pp. 87–112.
Capurro, Rafael. 2001. Einführung in die Informationswissenschaft. 195 pp. [http://www.
capurro.de/iwmodul1.html (last visit: 26.10.2006)]
Dehmer, matthias. 2005. ‘Data mining-Konzepte und graphentheoretische methoden
zur Analyse hypertextueller Daten.’ In: mehler/wolff. 2005a. pp. 113–141.
Dörre, Jochen; Gerstl, Peter; Seiffert, Roland. 2004. ‘Volltextsuche und text mining.’ In:
Carstensen, Kai-Uwe; Ebert, Christian; Endriss, Cornelia; Jekat, Susanne; Klabunde,
Ralf; Langer, hagen. 2004. Computerlinguistik und Sprachtechnologie. Eine Einführung.
2., überarbeitete und erweiterte Aulage. heidelberg: Elsevier Gmbh. ISBN 3-8274-
1407-5. xiv + 642 pp. pp. 479–495.
[the contribution to this manual is supposed to have an introductory approach.
Unfortunately there are not any illustrating examples which could help to verify
the different learning steps.]
Dunbill, Edd. 2000. The semantic web: a primer. 8 pp. [http://www.xml.com/lpt/a/653 (last
visit: 24.10.2006)]
ferber, Reginald. 2003. Information Retrieval. Suchmodelle und Data-Mining-Verfahren für
Textsammlungen und das Web. heidelberg: dpunkt-Verlag. ISBN 3-89864-213-5. xi +
340 pp.
forster, Julian. 2005. Die Transformation von Text in Vektoren. 11 pp. [http://www.informatik.
uni-ulm.de/ni/Lehre/SS05/Proseminartextmining/ausarbeitungen/forster.pdf (last
visit: 6.9.2006)]
fritschi, michael. 2003. ‘text-mining bring Licht in die Daten-Gerümpelkammer.’ In:
Netzwoche 21. p. 22. [http:www.spss.ch/upload/1057821904_Artikel%20Netzwo
che%20zu%20text%20mining.pdf (last visit: 5.9.2006)]
01_2007_5222_txt_ML.indd 166 6-12-2007 15:14:07

WORKSHOP
fuhr, Norbert. 2004. Information Retrieval. Skriptum zur Vorlesung im SS 04. [http://www.
is.informatik.uni-duisburg.de/courses/ir_SS04/folien/irskall.pdf (last visit: 6.9.2006)]
hausser, Roland. 2006. A computational model of natural language communication. Interpretation,
inference, and production in database semantics. Berlin; heidelberg; New york:
Springer. ISBN 978-3-540-35476-5; 3-540-35476-X. xii + 365 pp.
hearst, marti A. 1999. Untangling Text Data Mining. [http://www.ischool.berkeley.edu/
~hearst/papers/acl99/acl99-tdm.html (last visit: 30.7.2007)]
heyer, Gerhard; quastfoff, Uwe; wittig, thomas. 2006. Text Mining: Wissensrohstoff Text.
Konzepte, Algorithmen, Ergebnisse. herdecke, Bonn: w3L-Verlag. ISBN 3-937137-
30-0. 348 pp.
hippner, hajo; Rentzmann, René. 2006. ‘text mining.’ In: ‘Gesellschaft für Informatik
e.V. Informatik Lexikon. [http://www.gi-ev.de/service/informatiklexikon/informatik
lexikon-detailansicht/meldung/137/ (last visit: 6.9.2006)]]
hofherr, ferdinand. 2005. Information Retrieval. 11 pp. [http://www.informatik.uni-ulm.
de/ni/Lehre/SS05/Proseminartextmining/ausarbeitungen/hofherr.pdf (last visit:
6.9.2006)]
honsel, Gregor. 2006. Text Mining für das Internet. [http://www.heise.de/tr/artikel/75239
(last visit: 6.9.2006)]
hotho, Andreas; maedche, Alexander; Staab, Steffen; Zacharias, Valentin. 2002. On
knowledgeable supervised text mining. 22 pp. [http://www.fzi.de/KCmS/kcms_ile.
php?action=link &id=44 (last visit: 6.9.2006)]
hotho, Andreas; Nürnberger, Andreas; Paaß, Gerhard. 2005. ‘A brief survey of text mining.’
In: mehler/wolff. 2005a. pp. 19–62.
hotho, Andreas; Staab, Steffen; Stumme, Gerd. 2003. Text clustering based on background
knowledge. 41 pp. [http://www.aifb.uni-karlsruhe.de/~sst/Research/Publications/
text-mining-tech-report-425.pdf (last visit: 5.9.2006)]
hutchins, John. 1986. Machine translation: past, present, future. Chichester (UK): Ellis horwood.
ISBN 0-85312-788-3. 382 pp. [Electronic version: http://www. hutchinsweb.me.uk/PPf-tOC.htm
(last visit: 30.7.2007)]
Ihlenfeld, manuel. 2005. Neuronale Netze zur Dokumentenklassifizierung. 13 pp. [http://
www.informatik.uni-ulm.de/ni/Lehre/SS05/Proseminartextmining/ausarbeitungen/
Ihlenfeld.pdf (last visit: 6.9.2006)]
Jochim, maxim. 2006. Verfahren des Information Retrieval. Seminararbeit. Universität Karlsruhe.
14 pp. [http://www.ipd.uni-karlsruhe.de/~oosem/S2D2/material/3-Jochimm.
pdf (last visit: 25.10.2006)]
Kao, Anne; Poteet, Steve R. (eds). 2006. Natural language processing and text mining. London:
Springer. ISBN 978-1-84628-175-4. xii + 265 pp.
166 | 167
01_2007_5222_txt_ML.indd 167 6-12-2007 15:14:07

Kirchner, marc. 2002. Semantische Netze. 17 pp. [http://www.linuxco.de/stud/kirchner_
osem_semnet_ausarbeitung.pdf (last visit: 26.10.2006)]
Kogan, Jacob (ed.). 2006. Grouping multidimensional data: recent advances in clustering. Berlin;
heidelberg; New york: Springer. ISBN 978-3-540-28348-5; 3-540-28348-X.
xii + 268 pp.
Krause, michael. 2007. Kontextbereitstellung in offenen, ubiquitären Systemen. Dissertation,
Universität münchen 2006. münchen: Verlag Dr hut. ISBN 978-3-89963-469-3;
3-89963-469-1. x + 204 pp.
Kunz, Christoph. s.d. Semantics-driven information retrieval: an integrated approach. 10 pp.
[http://www.hci/iao/fraunhofer.de/uploads/tx-publications/050427_semantics_
driven_ir_v1.2.pdf (last visit: 25.10.2006)]
Kunz, Christoph. 2006. Ein integrierter Ansatz zur wissensbasierten Informationsrecherche. Dissertation
Universität Stuttgart. 139 pp. [http://deposit.ddb.de/cgi-bin/dokser?idn=979977142
&doc_var=d1 &doc_ext=pdf &ilename=979977142. pdf (last visit: 24.10.2006)]
Leopold, Edda. 2005. ‘On semantic spaces.’ In: mehler/wolff. 2005a. pp. 63–86.
Liebman, michael N. 2003. ‘Digging for nuggets of wisdom with text mining’, The New
York Times, 16 October 2003. [http://www.spss.ch/upload/1122650513_Liebman%
20Digging%20for%20Nuggets%20of%20wisdom%20with%20text%20mining%
20.pdf (last visit: 5.9.2006)]
mehler, Alexander; Köhler, Reinhard. 2006. Aspects of automatic text analysis. (Studies in
fuzziness and soft computing 209). Berlin; heidelberg; New york: Springer. ISBN
978-3-540-37520-3; 3-540-37520-1. ix + 464 pp.
mehler, Alexander; Lobin, henning (eds). 2004. Automatische Textanalyse. Systeme und
Methoden zur Annotation und Analyse natürlichsprachiger Texte. wiesbaden: Verlag für
Sozialwissenschaften. ISBN 3-531-14181-3. viii + 290 pp.
mehler, Alexander; wolff, Christian (eds). 2005a. Text mining. (LDV forum, Zeitschrift
für Computerlinguistik und Sprachtechnologie, GLDV-Journal for Computational
Linguistics and Language technology 20, 1). ISSN 0175-1336. v + 143 pp. [http://
www.uni-regensburg.de/fakultaeten/phil_fak_IV/medieninformatik/
LDV_forum_Band_20_heft_1.pdf (last visit: 5.9.2006)]
mehler, Alexander; wolff, Christian. 2005b. ‘Perspektiven und Positionen des text mining.’
In: mehler/wolff. 2005a. pp. 1–18.
Nohr, holger. 2001. Automatische Indexierung. Einführung in betriebliche Verfahren, Systeme
und Anwendungen. (materialien zur Information und Dokumentation 13).
Potsdam: Verlag für Berlin-Brandenburg. ISBN 3-935035-19-5. 108 pp.
Nohr, holger. 2003. Grundlagen der automatischen Indexierung. Ein Lehrbuch. Berlin:
Logos-Verlag. ISBN 3-8325-0121-5. 153 pp.
01_2007_5222_txt_ML.indd 168 6-12-2007 15:14:07

WORKSHOP
Oberle, Daniel; Staab, Steffen; Studer, Rudi; Volz, Raphael. 2003. Supporting application
development in the semantic web. 31 pp. [http://www.aifb.uni-karlsruhe.de/
~sst/Research/Publications/acm-toit-kaon-server.pdf (last visit: 5.9.2006)]
Paulsen, Christina. 2005. Extracting named entities from documents. 15 pp. [http://www.
informatik.uni-ulm.de/ni/Lehre/SS05/Proseminartextmining/ausarbeitungen/
Paulsen.pdf (last visit: 6.9.2006)]
Piazzi, Kai-Uwe. 2005. Digital Libraries and Textmining. 12 pp. [http://www. informatik.
uni-ulm.de/ni/ Lehre/SS05/Proseminartextmining/ ausarbeitungen/Piazzi.pdf (last
visit: 6.9.2006)]
punkt. netServices. s.d. Text-mining. [http://www.punkt.at/text.php?main=2&sub=3&id
=7 (last visit: 6.9.2006)]
Reich, Björn. 2005. Lineare und Probalistische Dokumentenklassifizierung. 13 p. [http: //www.
informatik.uni-ulm.de/ni/Lehre/SS05/Proseminartextmining/ausarbeitungen/
Reich.pdf (last visit: 6.9.2006)]
Robb, Drew. 2004. ‘text mining tools take on unstructured data’, Computerworld Business
Intelligence, 21 June 2004. [http://www.computerworld.com/action/article.do?com
mand=viewArticletOC&specialReportId=560&articleId=93968 (last visit:
22.10.2006)]
Schweighofer, Erich. 1999. Rechtsinformatik und Wissensrepräsentation. Automatische Textanalyse
im Völkerrecht und Europarecht. (forschungen aus Staat und Recht 124). wien;
New york: Springer. ISBN 3-211-83216-5. xx + 440 pp.
Semio Corporation. 2006. ‘text mining and the knowledge management space’. white
Paper. [http://www.dmreview.com/whitepaper/paper_sub.cfm?whitepaperId=
10080 (last visit: 6.9.2006)]
Stern, matthias. 2004. Verwendung von Ontologien zur Verbesserung von Informationsgewinn in
E-Learningsystemen. Diplomarbeit. technische Universität Graz. 144 pp. [http://
www.iicm.tu-granz.at/thesis/mstern.pdf (last visit: 24.10.2006)]
Studer, Rudi; Oppermann, hendrik; Schnurr, hans-Peter. 2001. Die Bedeutung von Ontologien
für das Wissensmanagement. (Ontoprise — Semantics for the web). 7 pp.
[http://www.ontoprise.de/documents/Bedeutung_von_Ontologien_fuer_wm.pdf
(last visit: 25.10.2006)]
Studer, Rudi; Schnurr, hans-Peter; Nierlich, Andreas. s.d. Semantik für die nächste Generation
Wissensmanagement. 20 pp. [http://www.community-of-knowledge.de/pdf/f05.
pdf (last visit: 25.10.2006)]
Studer, Rudi; Schnurr, hans-Peter; Nierlich, Andreas. 2001. Semantisches Knowledge Retrieval.
(Ontoprise — Semantics for the web). 15 pp. [http://www.ontoprise.de/documents/
Semantisches_Knowledge_Retrieval_whitepaper.pdf (last visit: 25.10.2006)]
168 | 169
01_2007_5222_txt_ML.indd 169 6-12-2007 15:14:07

Vernay, henri. 1991. Dictionnaire onomasiologique des langues romanes. tübingen: Niemeyer.
ISBN 3-484-50321-1. 6 vols.
Volz, Raphael; handschuh, Siegfried; Staab, Steffen; Stojanovic, Ljiljana; Stojanovic,
Nenad. 2004. ‘Unveiling the hidden bride: deep annotation for mapping and migrating
legacy data to the semantic web’. In: Web semantics: science, services and agents
on the world wide web 1. pp. 187–206. [http://www.aifb.uni-karlsruhe.de/~sst/
Research/Publications/2004/inal-hideen-bride.pdf. (last visit: 5.9.2006)]
werner, Lars. 2007. Typografiegewichtete Information-Retrieval-Verfahren in Dokumentenmanagementsystemen.
(C-LAB publication 23). Dissertation, Universität Paderborn 2006.
Aachen: Shaker. ISBN 978-3-8322-6054-5; 3-8322-6054-4. iv + 132 pp.
01_2007_5222_txt_ML.indd 170 6-12-2007 15:14:08

En tant que déléguée de la Grèce au groupe de travail “Informatique juridique”
et ayant suivi régulièrement les travaux de ce groupe au sein du Conseil
des ministres, je me réjouis d’avoir la possibilité de fêter les 25 ans du droit
communautaire en ligne et plus de 25 ans d’existence de ce groupe et surtout
de son travail fructueux et eficace. PRESS REVIEW
C’était en 1983 que mon pays, la Grèce, se réjouissait du nouveau système
Celex, prometteur et qui nous avait REVUE ébloui. J’écrivais DE alors PRESSE
à l’époque, dans une
revue juridique grecque , combien il était nécessaire de disposer d’un système
donnant accès à toute l’information ju ridique et judicaire, tant pour les services
publics et les universités que pour les particuliers qui s’intéressent au droit
et à son application. Cette nécessité de recherche documentée est devenue
encore plus aiguë en droit communautaire vu le nombre d’actes adoptés, la
diversité des secteurs couverts, la nature spéciale de l’ordre juridique de chaque
état membre, ainsi que la diversité des langues oficielles.
Par sa résolution du 26 novembre 1974 concernant la documentation juridique
automatique, le Conseil des ministres de la Justice avait lancé l’idée de
la création d’un système communautaire de documentation juridique automatique
ouvert à tous les états membres, puisque ceux-ci devaient appliquer le
droit communautaire.
Dans l’article, je me référais aussi au champ couvert par le système, qui
concernait non seulement les textes des traités, mais également le droit dérivé,
les accords conclus par l’Union européenne (à l’époque la Communauté européenne),
la jurisprudence de la Cour de justice ou encore les actes préparatoires.
Ce qui était étonnant, c’était que le système s’étendait aussi aux actes
adoptés par les états membres pour mettre en œuvre le droit communautaire
au sein de leur ordre juridique interne et mentionnait déjà la jurisprudence
nationale. Le système incluait des actes depuis 1952 et la mise à jour se faisait
avec un délai de 3 à 4 semaines à partir de la publication de l’acte. Chaque institution
communautaire était chargée des secteurs le concernant. Le Conseil
fournissait ainsi des données concernant les traités, les relations extérieures, le
droit dérivé, le droit complémentaire. La Commission, le Parlement européen
et le Comité économique et social fournissaient les actes préparatoires. La
Cour fournissait la jurisprudence et le Parlement les questions parlementaires.
La procédure de chargement débutait par un formulaire en deux parties.
Une partie, assez détaillée, comprenait plus de 30 rubriques de données documentaires,
et une deuxième présentait le texte entier ou un résumé de celui-ci,
selon le schéma du formulaire. On pourrait mentionner certains types de ru-
01_2007_5222_txt_ML.indd 171 6-12-2007 15:14:08

"
B.I.T.online
Zeitschrift für Bibliothek,
Information und Technologie
mit aktueller Internet-Präsenz
10 (2007) AUSGABE 1
01_2007_5222_txt_ML.indd 174 6-12-2007 15:14:18

PRESS REVIEW / REVUE DE PRESSE
"
174 | 175
01_2007_5222_txt_ML.indd 175 6-12-2007 15:14:21

01_2007_5222_txt_ML.indd 176 6-12-2007 15:14:25

01_2007_5222_txt_ML.indd 178 6-12-2007 15:14:25

European Commission
25 years of European law online — The event
25 années de droit européen en ligne — L’événement
Luxembourg: Ofice for Oficial Publications of the European Communities
2007 — 177 pp. — 20 × 24 cm
ISBN 978-92-78-40413-0
Price (excluding VAt) in Luxembourg: EUR 25
01_2007_5222_txt_ML.indd 179 6-12-2007 15:14:25

01_2007_5222_txt_ML.indd 180 6-12-2007 15:14:26