02.03.2013 Views

SDI Convergence - Nederlandse Commissie voor Geodesie - KNAW

SDI Convergence - Nederlandse Commissie voor Geodesie - KNAW

SDI Convergence - Nederlandse Commissie voor Geodesie - KNAW

SHOW MORE
SHOW LESS

You also want an ePaper? Increase the reach of your titles

YUMPU automatically turns print PDFs into web optimized ePapers that Google loves.

”Member States may limit public access to spatial data sets and services [...] where<br />

such access would adversely affect [...] the protection of the environment to which such<br />

information relates, such as the location of rare species.”<br />

We argue here that map-layers represented in static ways powered by static legal- and<br />

policy constraints do exactly that. They do not support conflict resolution and negotiation,<br />

but rather suggest more inflexibility and legal rigidity. INSPIRE based static map<br />

layers may be counterproductive to conflict resolution because of the tendency to be<br />

too general or too specific with no dynamic adjustment possibilities based on flexible<br />

regulative parameters. The interesting difference between a screen showing map layers<br />

already there and those that ‘turn up’ while moving a qualified cursor (like the one<br />

seeking space for open sailing areas) is the affordance of opportunity finding in contrast<br />

with the annotation of an area that is ‘locked up’. The provinces call this functionality<br />

a ‘seeking area’. They have created the unusual legal term ‘seeking area’ to obtain<br />

legal degrees of freedom in development plans that do not occur with fixed parameterisation.<br />

Feedback on the tests with the Flevoland regional development plan prove that questions<br />

like: return all contours on the map that fulfils the legal constraints ‘X’ and ‘Y’, but<br />

not ‘Z’ are answerable. How does one provide such type of opportunity finding for the<br />

user in a meaningful representation that allows for more flexibility? We have argued<br />

that INSPIRE Maps showing Natura2000 areas or sites should enable the functionality<br />

of ‘seeking area’. Simcity game developers who created manoeuvrability using a cursor<br />

and ‘tiles’ with fixed business rules may have developed the answer already. This was<br />

in 1985 when Simcity was still called ‘Micropolis’ (Wright, 2004a).<br />

6. CONFLICT ANNOTATION ENGINE FOR LEGAL ATLAS III<br />

Simcity-like functionality (see Wright, 2004b) resembles the required flexibility or ‘seeking<br />

area’. It is mostly based on Semantic Web Technology. The knowledge models<br />

about the legal constraints and the domain knowledge of the working scenarios are all<br />

described with Resource Description Framework/ Web Ontology Language (RDF/<br />

OWL). We chose RDF/OWL to infer and reason with these models. To publish this information<br />

as a service we use the OpenRDF Sesame server. This server has an<br />

SPARQL (SPARQL Protocol and RDF Query Language)-endpoint, which is an access<br />

point to which SPARQL queries can be sent. The SPARQL-endpoint is accessible<br />

through the web. The RDF that is stored within the OpenRDF Sesame server is processed<br />

with OWLIM. OWLIM is a high-performance semantic repository. It is packaged<br />

as a Storage and Inference Layer (SAIL) for the Sesame RDF database. It reasons<br />

about the RDF data and propagates this by means of rule-entailment. The SPARQLendpoint<br />

is used to fill the Legal Atlas III with information. The Legal Atlas III is as an<br />

interface for the OpenRDF Sesame server, and the SPARQL-endpoint is the interface<br />

between them.<br />

The SPARQL queries are based on the schemata of the RDF/OWL models (van de<br />

Ven et al., 2007). This means that they are independent on the content. This ensures<br />

that different content is annotated with the RDF/OWL models to ensure that the<br />

SPARQL queries are able to retrieve the content. The return is a gigantic list of all the<br />

concepts that can be used for annotation. Such a huge list might not be convenient in a<br />

user interface. Therefore it is better to replace this list with a practical list. Pruning this<br />

list down to a domain is one way to limit the amount of concepts. The following example<br />

shows the SPARQL query for a specific domain, namely the IMRO2006 (Informatiemodel<br />

Ruimtelijke Ordening, Dutch information model for spatial planning) SKOS<br />

68

Hooray! Your file is uploaded and ready to be published.

Saved successfully!

Ooh no, something went wrong!