- From: Charles McCathieNevile <charles@w3.org>
- Date: Tue, 16 Mar 2004 07:36:13 -0500 (EST)
- To: "Miles, AJ (Alistair) " <A.J.Miles@rl.ac.uk>
- Cc: "'Tudhope D S (Comp)'" <dstudhope@glam.ac.uk>, "'public-esw-thes@w3.org'" <public-esw-thes@w3.org>
Hi Alistair, a minor comment and a question. Comment - since rdfs:label is a human readable label rather than for machines, it seems better to me if you just put spaces between words. If you want to auto-generate property names from labels you can do so by globbing them together (or camel-casing them, or whatever) but as far as the machine is concerned you could also do so by selecting random combinations of arabic and chinese characters that don't already appear in your schema. This would probably encourage people to build interfaces that look for human-readable text to present things, so may be beneficial (although I suspect there are still enough hand-coders out there that there is some value in legible tag names). The question (because I don't know the answer) is "how would you express the disjoint condition in RDF?" cheers Chaals On Tue, 16 Mar 2004, Miles, AJ (Alistair) wrote: >Hi Doug, all, > >Should have mentioned, I tentatively decided to drop the skos:Facet class >and the skos:inFacet and skos:facetMember properties in SKOS-Core 1.0 after >there was some contention as to whether this had been modelled in the right >way. > >But Doug if you think the way it was done was OK for now, then I'm happy to >re-include it. > [snip] > >The statements above are based on the principal that facets are disjoint >classes (although the disjoint condition is sometimes broken and therefore >is not expressed as a formal constraint), and facet members are class >instances. > >Al. >
Received on Tuesday, 16 March 2004 07:36:14 UTC