W3C home > Mailing lists > Public > public-esw-thes@w3.org > November 2004

RE: inScheme v. OWL set semantics

From: Miles, AJ (Alistair) <A.J.Miles@rl.ac.uk>
Date: Mon, 15 Nov 2004 17:59:43 -0000
Message-ID: <350DC7048372D31197F200902773DF4C05E50D3F@exchange11.rl.ac.uk>
To: 'Jason Cupp' <jcupp@esri.com>, "'public-esw-thes@w3.org'" <public-esw-thes@w3.org>

Hi Jason,

> 
> If you create a concept scheme with concepts and make it public, you
> wouldn't want anyone to arbitrarly assert:
> <http://myconcept> <skos:inScheme> <http://publishedScheme> .
> 
> But theres no way in SKOS to describe a class (a skos:ConceptScheme),
> completly specified with an enumeration of it's members 
> (skos:Concept). 
> 
> An enumerated class in OWL seems to be a better way to say that these
> concepts belongs to this vocabulary. This would obsolete the 
> <skos:inScheme>
> property and give a catalog the knowledge of scheme and 
> concepts clearly
> defined in one place -- in the schema and not among the instances.
> 

This is worth further discussion I think.

What's to stop someone publishing an alternative enumeration?  Won't you
still need provenance data if someone tries to make spurious claims,
whichever mechanism you use?

Also, a concept scheme may not be simply an enumeration of a set of concepts
- most people from the thesaurus community consider a concept scheme as
consisting also of a set of semantic relationships between concepts. 

Am open to discussion and alternatives here, and as a general statement
support the re-use features of existing vocabs such as OWL, wherever
appropriate.

Al.



> - Jason
> 
Received on Monday, 15 November 2004 18:00:16 GMT

This archive was generated by hypermail 2.2.0+W3C-0.50 : Monday, 7 December 2009 10:38:52 GMT