The Cassetin Project — Towards an Inventory of Ancient ... - TUG
The Cassetin Project — Towards an Inventory of Ancient ... - TUG
The Cassetin Project — Towards an Inventory of Ancient ... - TUG
You also want an ePaper? Increase the reach of your titles
YUMPU automatically turns print PDFs into web optimized ePapers that Google loves.
<strong>The</strong> <strong>Cassetin</strong> <strong>Project</strong>—<strong>Towards</strong> <strong>an</strong> <strong>Inventory</strong> <strong>of</strong> <strong>Ancient</strong> Types<br />
<strong>an</strong>d the Related St<strong>an</strong>dardised Encoding<br />
Jacques André<br />
Irisa/Inria-Rennes<br />
Campus de Beaulieu<br />
F-35042 Rennes Cedex, Fr<strong>an</strong>ce<br />
Jacques.Andre@irisa.fr<br />
Introduction<br />
Abstract<br />
<strong>The</strong> <strong>Cassetin</strong> <strong>Project</strong> purposes are to inventory all <strong>of</strong> the types used in Europe for centuries, to<br />
st<strong>an</strong>dardize their encoding or naming <strong>an</strong>d to propose this list as a Unicode addendum (or at least<br />
as aprivate area).<br />
Résumé<br />
Nousdécrivonsleprojet<strong>Cassetin</strong>d’inventairedestypesutilisésentypographieeuropéenneenvue<br />
de la normalisation de leur codage ou de leur nommage, voire de leur intégration d<strong>an</strong>s Unicode<br />
(au pire d<strong>an</strong>s unezone privée).<br />
CharacterencodingssuchasUnicode[15]makeastrong<br />
difference between characters (abstract linguistic entities)<br />
<strong>an</strong>d glyphs (the rendition <strong>of</strong> characters) <strong>an</strong>d so ignorethe“types”effectivelyusedwhencomposingbooks.<br />
Typical examples are ligatures (such as “�”), abbreviations<br />
orv<strong>an</strong>ishedcharacters (suchas theoldFrench“˛c ˛e<br />
˛n”). This absence makes it difficult to st<strong>an</strong>dardize OCR<br />
outputs <strong>an</strong>d quite impossible to get genuine plain text<br />
from electroniceditions<strong>of</strong>old books(especiallyRenaiss<strong>an</strong>ce<br />
or even 18th century ones). Str<strong>an</strong>gely, far more<br />
complicatedtexts,suchasmedievalm<strong>an</strong>uscripts,arenow<br />
electronicallyeditableth<strong>an</strong>ks to encodingprojects.<br />
We first show how encodings have been applied to<br />
medievalm<strong>an</strong>uscripts,thendescribetheequivalent,<strong>an</strong>d<br />
not favorable, situation for old books. We follow with a<br />
presentation <strong>of</strong> the <strong>Cassetin</strong> project: its aims are to inventory<br />
all <strong>of</strong> the Europe<strong>an</strong> types, to name them <strong>an</strong>d to<br />
propose this listas a Unicodec<strong>an</strong>didate.<br />
Electronic Editing <strong>of</strong> M<strong>an</strong>uscripts<br />
More<strong>an</strong>dmorem<strong>an</strong>uscriptsareeditedontheWeboron<br />
CDs: medievalcharts,writers’drafts,registers<strong>of</strong>births,<br />
etc. Most <strong>of</strong> them are only available in image mode.<br />
Some are accomp<strong>an</strong>ied with the corresponding text, or<br />
ratherwiththecorresponding“texts”whentheeditionis<br />
undertakenbyhum<strong>an</strong>itiesscholars. Indeedtheyaredifferent<br />
views<strong>of</strong> the same text.<br />
Let us take as <strong>an</strong> example a m<strong>an</strong>uscript <strong>of</strong> Bernard<br />
Fig. 1: M<strong>an</strong>uscript <strong>of</strong> De Ventadour (Fr<strong>an</strong>ce, 12th<br />
century)<br />
[...] ch<strong>an</strong>ter et vint conter et enseigner.<br />
[...]tosing<strong>an</strong>d came singing<strong>an</strong>d teaching.<br />
2-a: Moderntr<strong>an</strong>slations<br />
[...]ch<strong>an</strong>tar et venc cortezet enseingnatez.<br />
2-b: Moderncomposition<br />
[...] chātar<br />
7 venccortef 7ēfeingnatz.<br />
2-c: Diplomatic edition<br />
Fig. 2: Various editions <strong>of</strong> figure 1 (last two lines <strong>of</strong><br />
text)<br />
314 <strong>TUG</strong>boat,Volume24 (2003), No. 3—Proceedings<strong>of</strong> EuroTEX2003
... ch&<strong>an</strong>;tar &et7; venc corte&longz; &et7;<br />
&en;&longs;eingnatz.<br />
Fig. 3: TEIencoding<strong>of</strong> figure1<br />
GLYPH MUFI ISO UNICODE<br />
ENTITY NAME ENTITY CODE NAME<br />
&con; sign con 0254 small open o<br />
✗ sign cross 271D latin cross<br />
&ed; sign ed ; 003D semicolon<br />
&est; sign est F150 homothetic<br />
&etslash; sign et with slash<br />
Fig. 4: Example <strong>of</strong> MUFIproposal for encodingNordic medievalabbreviations<br />
de Ventadour 1 <strong>an</strong>d even let us consider just the last two<br />
lines<strong>of</strong> Figure1. <strong>The</strong>yc<strong>an</strong> beencodedin variousways.<br />
In reverse order, we c<strong>an</strong> first tr<strong>an</strong>slate these lines into<br />
modern French or even into modern English (figure 2a).<br />
Astudentinmedievalphilologywouldpreferaform<br />
closer to the genuine l<strong>an</strong>guage, like figure 2-b, where<br />
each sign is composed with modern types, without respectto<br />
specifich<strong>an</strong>dwrittensigns. Away 2 closerto the<br />
m<strong>an</strong>uscript is to edit the text with types close to the<br />
h<strong>an</strong>dwritten signs (figure 2-c) <strong>an</strong>d to respect the layout<br />
(hyphenation,etc.).<br />
Intheend,aresearcherneedsfarmore,suchasligatures<br />
or signs used,unclear letters,who wrote or <strong>an</strong>notatedtheoriginaltext<strong>an</strong>dwhohastr<strong>an</strong>slatedorencoded<br />
it. M<strong>an</strong>yprojectshavebeenlaunchedtoaccesselectronic<br />
versions <strong>of</strong> medieval m<strong>an</strong>uscripts. 3 <strong>The</strong>y generally use<br />
SGML-like tags to indicate either the structure <strong>of</strong> the<br />
1. A French troubadour (12th Century). Here is <strong>an</strong> example<br />
<strong>of</strong> a version <strong>of</strong> his Qu<strong>an</strong>dveilalaudetamover... (When the skylark<br />
wings...), a song written in “Provencal” (<strong>an</strong> Old French dialect).<br />
2. This is close to the so called “diplomatic” version, generallyusedformodernwriters’drafts,thattakescareaswell<strong>of</strong>the<br />
position (orientation, length, etc.) <strong>of</strong> lines in pages.<br />
3. Such as Digital Scriptorium (http://sunsite.berkeley.<br />
edu/Scriptorium/), EAMMS (http://www.csbsju.edu/hmml/<br />
<strong>The</strong><strong>Cassetin</strong> <strong>Project</strong><br />
document or the actual characters used. Figure 3 shows<br />
how our two lines could be encoded. 4 Tags like &et7;<br />
or &<strong>an</strong>; indicate special characters (“et” <strong>an</strong>d “<strong>an</strong>”) <strong>an</strong>d<br />
are like the ones used by <strong>Project</strong> Charrette at Princeton<br />
University. 5 Otherprojectsuseothertags(e.g.,&abar;<br />
instead <strong>of</strong> &<strong>an</strong>;). That is why some new projects, such<br />
as MUFI 6 (figure 4), try to inventory all <strong>of</strong> these signs,<br />
to give them names (used as entity names), to propose 7<br />
themas Unicodecharacters <strong>an</strong>d finallyto propose a font<br />
<strong>of</strong>feringall <strong>of</strong> theseglyphs.<br />
From now on, th<strong>an</strong>ks to such projects, the various<br />
electronic editions <strong>of</strong> m<strong>an</strong>uscripts should follow a st<strong>an</strong>dardizedencoding,hencebecomeportable<strong>an</strong>dusableby<br />
<strong>an</strong>yresearcher.<br />
eamms/), or the TEI based project MASTER (http://www.cta.<br />
dmu.ac.uk/projects/master/).<br />
4. That form is rich: it c<strong>an</strong> be tr<strong>an</strong>slated to forms 2.a to 2.c,<br />
while the opposite way is generally impossible.<br />
5. <strong>The</strong> list <strong>of</strong> entities may be found at http://www.mshs.<br />
univ-poitiers.fr/cescm/l<strong>an</strong>celot/keys.html.<br />
6. Medieval Unicode Font Initiative, http://www.hit.uib.<br />
no/mufi/.<br />
7. Seesection“Unicode”belowfortheposition<strong>of</strong>theUnicode<br />
Consortium regarding glyphs <strong>an</strong>d characters.<br />
<strong>TUG</strong>boat,Volume24 (2003), No.3—Proceedings<strong>of</strong> EuroTEX2003 315
Jacques André<br />
thebiggest<strong>of</strong>hisLetterswithglasses,whichsomagnifiedthe<br />
whole Letter, that I could easily distinguish, <strong>an</strong>d with small<br />
dividersmeasure<strong>of</strong>fthesize,situation<strong>an</strong>dform<strong>of</strong>everypart,<br />
<strong>an</strong>dtheproportioneverypartboretothewhole;<strong>an</strong>dformy<br />
ownfuturesatisfactioncollectedmyobservationsintoabook,<br />
whichIhaveinsertedinmyExercisesonLetter-Cutting. For<br />
therein I have exhibitedto[...]<br />
5-a: Modern tr<strong>an</strong>sliteration<br />
the bigge&st; <strong>of</strong> his Letters with<br />
Gla&ss;es&thinspace;, which &longs;o<br />
magni&fi;&ed the whole Letter,<br />
that I could ea&longs;ily<br />
di‐&stlig;tingui&shlig;, [...]<br />
[...]<br />
&longs;atisfa&ctlig;ion collec&ctlig;ed [...]<br />
5-c: TEI-likeencoding<br />
Fig. 5: Extract from Moxon’s Mech<strong>an</strong>ick Exercises<br />
(1683), (volume 2 Applied to the Art <strong>of</strong> Printing)<br />
followed by a modern tr<strong>an</strong>sliteration (5-a) <strong>an</strong>d a<br />
TEI-like encoding (5-c)<br />
Electronic Editions <strong>of</strong> Old Books<br />
Probably because printed characters are thought to be<br />
more readable th<strong>an</strong> h<strong>an</strong>dwritten ones, quite <strong>of</strong>ten digitizedbooksareeditedasimages<strong>an</strong>d/oras<br />
plaintext encoded<br />
with modern characters. If some other types are<br />
used,no attempt ismade to st<strong>an</strong>dardizetheirencoding.<br />
Moxon’sMech<strong>an</strong>ickExercises Letusconsiderfirstthefamous<br />
Moxon’s book (1683, a pioneer in the matter <strong>of</strong><br />
printers’m<strong>an</strong>uals[4])Mech<strong>an</strong>ickExercises[14]. Figure5<br />
shows<strong>an</strong>extract<strong>an</strong>dthecorresponding“plaintext”version<br />
<strong>of</strong> it: it is a modern version, with modern words,<br />
no emphatic caps, no ligatures, etc. As for m<strong>an</strong>uscripts,<br />
it is possible to encode this text (figure 5-c) to take care<br />
<strong>of</strong> specific characters (like �), hyphenation (tag ÷),<br />
etc.<br />
However, m<strong>an</strong>y people would like <strong>an</strong> edition (like<br />
figure 6) that is both more legible th<strong>an</strong> TEI-encoding<br />
(figure 5-c) <strong>an</strong>d more authentic th<strong>an</strong> a traditional plain<br />
text. 8 Alas,ifonelooksinminutedetailattheseglyphs,<br />
onec<strong>an</strong>seethatifligaturessuchas“fi,”“longs+i,”“long<br />
s + t” or “ct” are present, other ones are absent, such as<br />
8. Furthermore,figure5-adoes notallow<strong>an</strong>yresearchon the<br />
use <strong>of</strong> the long s or <strong>of</strong> old words (e.g., scituation).<br />
the biggeſt<strong>of</strong> his...<br />
magnifiedthe whole...<br />
diſtinguish,<strong>an</strong>d withſmall...<br />
�ze,ſcituation...<br />
...<br />
own future ſatisfaſtion...<br />
Fig. 6: Moxon text (figure 5) edited with<br />
convenient glyphs 9<br />
Fig. 7: Fertel lowercase (1723)<br />
“sh”: itisquiteimpossiblet<strong>of</strong>indafontwithalltheusual<br />
(old)glyphs! Itisnotamatter<strong>of</strong>difficultytodesignthese<br />
characters,butsimplythatdesignersdonot knowthem!<br />
Notethattoday OCRsstillhaveproblemsrecognizing<br />
old types; however, current research allows one to<br />
think that very soon OCRs will be able to recognize Renaiss<strong>an</strong>ce<br />
characters <strong>an</strong>d their ligatures. What will be<br />
the output when recognizing the ligature “�”? A code<br />
(which one? nnn8, xxxx16?) or a name? Will it be<br />
“ct”(ifso,<strong>an</strong>ystudyonligatureswillbeimpossible)? Or<br />
“&ctlig;”? Or“&ct;”,or“\ct”,or...? St<strong>an</strong>dardized<br />
naming or encodingis required!<br />
Another example: Fertel’s case Figure 7 exhibits the<br />
lower case used by Fertel [8, page 12] in the first quarter<strong>of</strong>the18th<br />
century. Amongotherspecialcharacters<br />
(suchas ligatures“si”, “ssi” or “ct” <strong>an</strong>d typically French<br />
characters like “ç” <strong>an</strong>d “œ”), it <strong>of</strong>fers <strong>an</strong> “˛e” (left upper<br />
corner)thathasnothingtodowiththe“eogonek”used<br />
in eastern Europe. It is <strong>an</strong> echo <strong>of</strong> a former system to<br />
write French (by a grammari<strong>an</strong> <strong>an</strong>d poet, Je<strong>an</strong>-Antoine<br />
de Baïf, 16th century) <strong>an</strong>d this kind <strong>of</strong> “breve e” was<br />
usedforlessth<strong>an</strong>acentury. 10 Obviously,thischaracteris<br />
not<strong>an</strong>isolatedexample<strong>an</strong>dyouc<strong>an</strong>findm<strong>an</strong>ycharacters<br />
thatareneitherintoday’sencodingsnorinfontcharacter<br />
9. Here, as in the rest <strong>of</strong> these proceedings, the HW Caslon<br />
font is used, designed by Howard Jones.<br />
10. 50yearsafterFertel,Diderot’sEncyclopédieshowsthesame<br />
case,howeverthe“˛e”isreplacedby<strong>an</strong>“é”whilethegenuinebox<br />
for “é” is empty!<br />
316 <strong>TUG</strong>boat,Volume24 (2003), No. 3—Proceedings<strong>of</strong> EuroTEX2003
sets. Yet, they are present in foundry specimens, books<br />
or evenin grammars ...<br />
Unicode, Characters <strong>an</strong>d Glyphs<br />
Unicode [15] makes a strong difference between characters<br />
(abstract linguistic entities) <strong>an</strong>d glyphs (a possible<br />
physicalstylisticrepresentationorrendition<strong>of</strong>theseentities).<br />
Veryfewcleverpapersgiveagoodexpl<strong>an</strong>ation<strong>of</strong><br />
thoseconcepts;letuscitehereonebyKenWhistler,the<br />
technical director <strong>of</strong> Unicode [2] <strong>an</strong>d one by a typographer,<br />
John Hudson [12]. On the other h<strong>an</strong>d, there are<br />
also good papers that say that Unicode made the wrong<br />
choice <strong>an</strong>d that characters <strong>an</strong>d glyphs are not so easily<br />
different [10, 11]. We would like to add that “types”<br />
(with the usual typographic me<strong>an</strong>ing) are neither characters,<br />
nor glyphs. 11<br />
An import<strong>an</strong>t point is that the Unicode principle<br />
thatseparatesglyphs<strong>an</strong>dcharactershasbeenhistorically<br />
violated by <strong>an</strong>other one: Unicode is based on previous<br />
encodingsystems(proprietaryorinternationalst<strong>an</strong>dards)<br />
whereligatureswerepresent. IfUnicodewascle<strong>an</strong>,even<br />
the sign “&” should not be there! However we c<strong>an</strong> be<br />
suspiciouswhy“long s”<strong>an</strong>deven“ligaturest”havebeen<br />
veryrecentlyadded<strong>an</strong>dnot “ligaturect”!<br />
Imaginethe dialog:<br />
–“HowcouldIdescribe 12 Fertel’scase(figure7)<strong>an</strong>dits<br />
˛eusingUnicode?” I ask.<br />
– Answer from Unicode specialist: “Use latin small<br />
letter e with ogonek, U+0119.”<br />
–“No,Isay,Fertel’scharacterisnotthatcharacter,there<br />
is the same glyph resembl<strong>an</strong>ce as with latin capital<br />
a <strong>an</strong>d greek capital alpha, but they are different<br />
characters <strong>an</strong>d Unicodeencodesthemseparately.”<br />
–“Whydon’tyouencodethischaracteraslettere<strong>an</strong>da<br />
combiningdiacriticogonek?”<br />
– “For it is not <strong>an</strong> ogonek,rather a kind<strong>of</strong> breve,”I<strong>an</strong>swer.<br />
– “OK,” he says, “your ˛e is a glyph <strong>of</strong> some latin<br />
small letter with breve.”<br />
Idisagree,it’snotthesamebreveastheoneusedbyFertelin<strong>an</strong>othercase:<br />
“ĕ”,soit’snotthesamecharacter...<br />
And now, if you look at the alphabet given by the same<br />
Baïf, you c<strong>an</strong> see <strong>an</strong> “a with raising tail” that is<br />
11. <strong>The</strong>re are m<strong>an</strong>y stylistic vari<strong>an</strong>ts <strong>of</strong> our “˛e”! On the<br />
otherh<strong>an</strong>d,Unicodespeaksaboutrendition<strong>of</strong>abstractcharacters.<br />
However, what about the other way: when sc<strong>an</strong>ning documents,<br />
printedcharactersexistbeforethecorresponding“abstract”character,<br />
they are not only images <strong>of</strong> abstract characters, they are<br />
charactersbythemselvesat<strong>an</strong>intermediarylevelbetweenglyphs<br />
<strong>an</strong>d linguistic entities.<br />
12. Even if “[t]he Unicode St<strong>an</strong>dard is explicitly not aimed at<br />
being a system for facsimile representation <strong>of</strong> text” [2], one may<br />
need to quote such a character. Actually, it is not only a Unicode<br />
problem!<br />
<strong>The</strong><strong>Cassetin</strong> <strong>Project</strong><br />
rather a nasal O (its place in the alphabet is just before<br />
thePletter). Let us restart thesame dialog...<br />
Lastpoint: Unicodeknowsoldl<strong>an</strong>guagessuchasthe<br />
Runes or Ogham. Why should it ignore old Europe<strong>an</strong><br />
l<strong>an</strong>guages<strong>an</strong>dtheir writingusedfor centuries?<br />
<strong>The</strong> <strong>Cassetin</strong> <strong>Project</strong><br />
BeinginvolvedindigitizationprojectssuchasFournier’s<br />
M<strong>an</strong>uel typographique, 13 I am continuously confronted<br />
with such problems <strong>of</strong> coding or naming old 14 characters.<br />
Discussionswithm<strong>an</strong>ypeopleinvolvedinsuchtasks<br />
pushed me recently to undertake a project 15 to inventory<br />
these types <strong>an</strong>d try to establish a st<strong>an</strong>dardized list<br />
<strong>of</strong> names or ... codes.<br />
Its main aims are:<br />
<strong>Inventory</strong><strong>of</strong>types Prepare<strong>an</strong> inventory<strong>of</strong> all types used<br />
intexts 16 printedin Europe<strong>an</strong> 17 l<strong>an</strong>guages.<br />
Typical characters are<br />
• Ligatures,suchastheonesalreadyquotedhere(sh,<br />
si,st,...) <strong>an</strong>dm<strong>an</strong>yotherones(liketheHungari<strong>an</strong><br />
gz...).<br />
• V<strong>an</strong>ished characters, such as the “˛e,” the tailed A,<br />
etc.<br />
• Accented characters (like the old Sp<strong>an</strong>ish conson<strong>an</strong>ts).<br />
• Abbreviations.<br />
• Special characters such as verset <strong>an</strong>d respons (these<br />
twoareinUnicode,butm<strong>an</strong>yotherspecialcharacters<br />
arenot).<br />
• Historical typographical characters 18 (that are not<br />
alreadyinUnicode)such as raisedletters.<br />
Thisinventoryisbased on<br />
• Previousstudies,suchas[3,4,5,7],includingWeb<br />
pagessuchas Bolton’son cases [6].<br />
• Specimenspublishedby foundries.<br />
• <strong>Ancient</strong>books.<br />
• <strong>The</strong> MUFI project for m<strong>an</strong>uscripts!<br />
13. LikeMoxon’s, afamous18thcenturybookontype-cutting<br />
<strong>an</strong>d typefounding. See [9, 13] <strong>an</strong>d http://www.irisa.fr/<br />
faqtypo/BiViTy.<br />
14. Old me<strong>an</strong>s here before DTP! A typical example is the use,<br />
stillcurrent in 1950, <strong>of</strong> theabbreviation “crossed K” thatrepresents<br />
the Breton “ker” occurring in m<strong>an</strong>y names.<br />
15. Temporarily called CASSETIN: “cassetin” is the French<br />
name <strong>of</strong> case boxes. It c<strong>an</strong> st<strong>an</strong>d for “CASSE Type encodINg”<br />
... See also [1].<br />
16. One problem not yet solved: should we consider all types,<br />
even the ones used outside <strong>of</strong> plain text, such as ornaments <strong>an</strong>d<br />
rules? I do not think so, however the limits are not yet fixed!<br />
17. Thisisagain<strong>an</strong>unsolvedquestion: Whichl<strong>an</strong>guagesdowe<br />
consider? Latinones? WhataboutCyrillic,Greek,Hebrew,Arabic,<br />
Syriac, etc.? Actually, today it is only a matter <strong>of</strong> specialists<br />
working in this project ...<br />
18. We do not dare to speak about small caps!<br />
<strong>TUG</strong>boat,Volume24 (2003), No.3—Proceedings<strong>of</strong> EuroTEX2003 317
Jacques André<br />
Naming <strong>an</strong>d Encoding Each identified type will enter a<br />
file with typical glyphs, usages, etc. <strong>an</strong>d a name will<br />
be given. This name will be usable as <strong>an</strong> XML entity<br />
(&xxx;), a TEX name (\xxx), <strong>an</strong> (Adobeor OpenType)<br />
glyphname, etc.<br />
Each type will as well be assigned a number: the<br />
Unicode number when it exists; if not, a new number<br />
thatcouldbe<strong>an</strong>entryinsomeUnicodeuserprivatearea.<br />
Note that these glyph/character names permit not<br />
only st<strong>an</strong>dardizing output from OCR, but as well a st<strong>an</strong>dardizedwaytotypeinthesespecialsignsforrendition.<br />
Experimental Font As MUFI does, <strong>an</strong> existing font<br />
shouldbeupgradedto<strong>of</strong>ferall<strong>of</strong>these,say,glyphs: editing“facsimile”textslikethoseinfigures5or6shouldbe<br />
easier! Again,it’snotonlyamatter<strong>of</strong>actualglyphs,but<br />
rather,tobest<strong>an</strong>dardised,amatter<strong>of</strong>tableencoding. In<br />
that way, TEX LMs or Ωcould begoodc<strong>an</strong>didates!<br />
Calendar Thisprojectisstillaprivateundertaking,however,I’dliketomakeit<strong>an</strong>internationalproject<br />
19 (with,<br />
e.g.,Europe<strong>an</strong> Union help).<br />
Aquickgl<strong>an</strong>ceatthealreadypublishedcasesshows<br />
that the number <strong>of</strong> new characters to inventory is not<br />
verylarge<strong>an</strong>dthelistshouldnotbeaslarge! Sothiswork<br />
shouldnot last years ...<br />
Conclusion<br />
Glyphs or not, characters or not, types belong to a class<br />
that is not recognized by Unicode. Histori<strong>an</strong>s <strong>of</strong> books,<br />
<strong>of</strong> l<strong>an</strong>guages, etc. do need a st<strong>an</strong>dardization <strong>of</strong> their<br />
names,even<strong>of</strong>theirencodings,insuchawaythattheincreasing<br />
number <strong>of</strong> sites <strong>of</strong>fering digitized books c<strong>an</strong> be<br />
researchedinaportable m<strong>an</strong>ner.<br />
References<br />
[1] AndréJ.(2003),“Numérisationetcodagedescaractères<br />
de livres <strong>an</strong>ciens”, Numérisation et patrimoine<br />
(B. Coüasnon éd.), special issue <strong>of</strong> Document<br />
numérique,vol.7,num.3–4,2003,pp.127–142.<br />
[2] Andries P. (2002), “Entretien avec Ken Whistler,<br />
directeurtechniqueduconsortiumUnicode”,Document<br />
numérique, vol. 6, 3-4, pp. 13–49. http:<br />
//hapax.iquebec.com/hapax/<br />
[3] Barber G. (1969), French Letterpress Printing. A<br />
list <strong>of</strong> French printing m<strong>an</strong>uals <strong>an</strong>d other texts in<br />
French bearing on the technique <strong>of</strong> letterpress printing,<br />
Occasional publication num. 5, Oxford BibliographicalSociety.<br />
19. One goal <strong>of</strong> this paper is a call to participation! Don’t hesitate<br />
to mail to the author.<br />
[4] Baudin F. (1994), L’effet Gutenberg, Éditions du<br />
Cercledela Librairie,Paris.<br />
[5] BigmoreF.C.<strong>an</strong>dWym<strong>an</strong>C.W.H.(1978),ABibliography<br />
<strong>of</strong> Printing with Notes & Illustrations,<br />
Holl<strong>an</strong>d Press Ltd (London) <strong>an</strong>d Oak Knoll Books<br />
(USA).<br />
[6] Bolton D. (2002), Typecases, <strong>The</strong> Alembic Press,<br />
http://members.aol.com/typecase/<br />
[7] Dreyfus J. (1972), Type specimen facsimiles, London.<br />
[8] Fertel M.D. (1723), Lasciencepratiquedel’imprimerie...,SaintOmer.FacsimilebyLibriseditions,<br />
1998.<br />
[9] Fournier le jeune (1764 et 1766), M<strong>an</strong>uel typographiqueutileauxgensdelettresetàceuxquiexercentlesdifférentespartiesdel’artetdel’imprimerie,<br />
Paris,chezBarbou, 2 tomes.<br />
[10] Haralambous Y. (2000), “Unicode, XML, TEI,<br />
Ω <strong>an</strong>d Scholarly Documents”, 16th International<br />
UnicodeConference,Amsterdam, 24 p.<br />
[11] HaralambousY.(2002),“Unicodeettypographie:<br />
unamourimpossible”,Documentnumérique,vol.6,<br />
n o 3-4, p.105–137.<br />
[12] Hudson J. (2002), “Unicode, from text to type”,<br />
L<strong>an</strong>guageCultureType—InternationalTypedesign<br />
intheAge<strong>of</strong>Unicode(Berry ed.),ATyPI/Graphis,<br />
pp.24–44.<br />
[13] Mosley J. <strong>an</strong>d Carter H. (1995), <strong>The</strong>M<strong>an</strong>uelTypographique<strong>of</strong>Pierre-SimonFournierlejeune,3volumes,<br />
Darmstadt.<br />
[14] Moxon J. (1683), Mech<strong>an</strong>ik Exercises or the Doctrine<strong>of</strong>H<strong>an</strong>dy-works,printedforJosephMoxonon<br />
the Westside <strong>of</strong> Fleet-ditch, at the Sign <strong>of</strong> Atlas.<br />
Misc.eds.byDavis<strong>an</strong>dCarter(Dover,1978),De<br />
Vinne (1886), Johnson <strong>an</strong>d Gibson (Oxford UniversityPress,1978),etc.Notethatthesecondvolume,AppliedtotheArt<strong>of</strong>Printing,ismoredifficult<br />
to find.<br />
[15] <strong>The</strong> Unicode Consortium (2003), <strong>The</strong> Unicode<br />
St<strong>an</strong>dard, Version 4.0, Addison-Wesley, ISBN 0-<br />
321-18578-1. See also http://www.unicode.<br />
org.<br />
318 <strong>TUG</strong>boat,Volume24 (2003), No. 3—Proceedings<strong>of</strong> EuroTEX2003