Re: SKOS editor

Hi Quentin,
> Hi all,
>
> I'm planning on developing a SKOS plugin for Protégé 4.0. 

Very interesting!

> I have attached to this email a screenshot of what I think would be useful. I believe that for this editor to be useful, it should make sure the information available is consistent with the SKOS model as described in [1]. 
>
> As I was designing the potential layout, I found a few interesting questions to which I would like comments from the working group.
> 1. I was wondering if the working group had defined a list of requirement for SKOS editor.
>   

Not that I'm aware of. One can hypothesize it should implement the 
semantics defined in the reference, e.g. prohibiting the creation of 
inconsistent models.

> 2. Top concept: in owl everything is owl:Thing, and I was wondering if there was an equivalent in SKOS. I was thinking skos:ConceptScheme would be the higher node in the hierarchy but then I read that a SKOS model could have more than one concept scheme. Is this correct?
>   

Yes. Furthermore, I think nothing prevents user to create SKOS models 
that do not group their concepts into concept schemes...
I would say that generally there is no "ground" concept in SKOS.

> 3. Hierarchy: I think that the user should only be given one of the hierarchy properties (i.e. either skos:narrower or skos:broader), but assigning the inverse in the background. I was also wondering if several concepts are broader than another concept, does that mean that the siblings are skos:related.
>
> For example,
> ex:MSc skos:broader ex:Student
> ex:PhD skos:broader ex:student
> Does that always entail?
> ex:MSc skos:related ex:PhD
>   

Not at all!  I've seen it sometimes, but I don't see why this should be 
always the case. On the contrary, RT links (in thesaurus standard) are 
especially useful to point users at other part of a hierarchy than the 
one the explored concept is situated in.

> 4. I was wondering if the working group had defined an extension for SKOS files (e.g. skos, rdfs) to be used.
>   

I also don't know. Personally I use .skos, but I never thought twice 
about it. ".rdfs" should not be OK as a pure SKOS model will likely 
contain no explicit RDFS class ".owl" could be used (when concept 
schemes are also considered as ontologies) and ".rdf" should be always 
possible.

Cheers,

Antoine
> Regards,
>
> Quentin
>
> [1] http://isegserv.itd.rl.ac.uk/cvs-public/~checkout~/skos/drafts/integrity.html?rev=1.7
>  
>
> ******************************************
> * Quentin H. Reul                        *
> * PhD Research Student                   *
> * Department of Computing Science        *
> * University of Aberdeen, King's College *
> * Room 238 in the Meston Building        *
> * ABERDEEN AB24 3UE                      *
> * Phone: +44 (0)1224 27 4485             *
> * http://www.csd.abdn.ac.uk/~qreul       *
> ******************************************
>   

Received on Tuesday, 5 February 2008 08:53:31 UTC