W3C home > Mailing lists > Public > public-vocabs@w3.org > October 2013

Re: SKOS for schema.org proposal for discussion

From: Antoine Isaac <aisaac@few.vu.nl>
Date: Sun, 20 Oct 2013 16:25:21 +0200
Message-ID: <5263E7D1.1070304@few.vu.nl>
To: "<public-vocabs@w3.org>" <public-vocabs@w3.org>
Hi,

Interesting that the topic has been stalled for one week, especially in the middle of a discussion on naming ;-). It looks like it will end like earlier SKOS threads, which also lead to discussion on the general strategy for schema.org or this list [1]...

OK, if applications need to publish or consume concept-level data, we can point them to RDFa+SKOS. But if some here prefers to use the schema.org namespace, we can't really say it's wrong. Especially when better-known ontologies have been already integrated into Schema.org. The discussion should have happened for FOAF and GR. And if it happens now, still, it should have a broader scope than SKOS!

I also hear the point that relying on SKOS-like data is less good than trying to categorize 'concepts', so that they fit various schema.org classes (Person, Place, etc). Again this debate has already happened, in a way.
If a good, clean ontologization of thesauri, folksonomies etc was possible (ie., if people had resources for it), then there wouldn't be any need for SKOS in the first place, in the Semantic Web / Linked Data ecosystem.
Besides the logical pitfalls of shoehorning SKOS data into OWL ontologies, there's the problem of raising the barrier to the use of data. A range of simple applications like the one Stéphanes has presented don't need fully-fleged ontologies, or, here, fine-grained instances of schema.org's 'concrete' classes.


To come back to the naming...
SKOS was partly designed to reflect the shift to 'traditional' term-based knowledge organization systems to more 'conceptual' ones (a shift examplified by more recent thesaurus standard). As Jean-Pierre said, the whole point is having string and terms masquerading as something more structured. Having skos:Concept mapped to a schema:Term or anything that prominently feature 'term' will be harmful in this respect.

"Topic" may be counter-intuitive for all the cases when the resources are not used as subjects of documents.

Using 'concept' does not seem so harmful to me, in fact. I don't see how the general schema.org users could possibly live and breath by early DL work and CommonKADS...
'EnumConcept' carries a meaning of ordered listing I'm not comfortable with. But if Enumeration has been already used without that sense in schema.org, it may well fly.

If you are really desperate for another one, how about 'category'?

Best,

Antoine

[1] http://lists.w3.org/Archives/Public/public-vocabs/2013Jan/0033.html
Received on Sunday, 20 October 2013 14:25:51 UTC

This archive was generated by hypermail 2.3.1 : Tuesday, 6 January 2015 21:29:32 UTC