Re: An introduction of LOD features of EUNIS

Søren Roug wrote:
> Hi lod-public,
>
> Peter de Vries alerted me to the presence of this mailing list, and since 
> you're currently discussing species modeling, I thought I would add my 2 
> cents.
>   

Welcome!
> I'm the maintainer of a site called EUNIS, which is used by the European 
> Commission to determine whether species, habitat types or sites need a change 
> in legislation and protection. There are about 200.000 species in the 
> database. We have over the last couple of months given it an overhaul and 
> added some linked data functionality. It is still a work in progress, and 
> we'll continue the improvements. The way we have implemented Linked Data is to 
> look at the accept header, and then either send text/html or 
> application/rdf+xml without a redirection. This means that for e.g. 
> http://eunis.eea.europa.eu/species/1038 the HTML and the RDF output is the 
> same URL.
>   

What you are saying is: I have a descriptor document that's available in 
HTML (default) and RDF/XML formats, with either format delivered to user 
agents via content negotiation.

But, I don't see an Identifier for the main Subject of your Descriptor 
docs. For instance,  in the HTML descriptor doc all I see currently is a 
literal value: Falco peregrinus. When looking at the RDF descriptor doc 
variant I see: <http://eunis.eea.europa.eu/species/1038> . For very 
little cost (re. tweaking your RDF data) you could just add "#this" to 
<http://eunis.eea.europa.eu/species/1038> making it: 
<http://eunis.eea.europa.eu/species/1038#this>, within your RDF and HTML 
descriptor docs.

To conclude: you need to unambiguously Name your Descriptor Document and 
Descriptor Document Subject using distinct Generic HTTP URIs.

Kingsley

> A note about our semantics. We're not using the predicate 
> skos:closeMatch like Pete. We have created two predicates.
> 1.       sameSynonym, which links a binomial name and author to the same 
> binomial name and author in the foreign database. (taking into account 
> different spellings and abbreviations). The purpose is to validate that our 
> name is used by at least one other database.
> 2.       sameSpecies, which links from a EUNIS accepted name to an accepted 
> name in the foreign database. The side-effect is that the species name might 
> change when you follow the link. sameSpecies is a sub-property of owl:sameAs.
>  
> We also have negative matches: notSameSynonym and notSameSpecies. These are 
> used when there is a high likelyhood of assuming it is the same species, and a 
> maintainer has determined it is not. 
>
> Practical examples:
>
> Danaus plexippus (Monarch butterfly) http://eunis.eea.europa.eu/species/90910
> Canis lupus (Gray wolf) http://eunis.eea.europa.eu/species/90910
> The Polish site Lasy Janowskie http://eunis.eea.europa.eu/species/90910
>
> Best regards,
>
> Søren Roug
> European Environment Agency
>
>
>   


-- 

Regards,

Kingsley Idehen	      
President & CEO 
OpenLink Software     
Web: http://www.openlinksw.com
Weblog: http://www.openlinksw.com/blog/~kidehen
Twitter/Identi.ca: kidehen

Received on Saturday, 5 June 2010 13:26:57 UTC