- From: <danbri@w3.org>
- Date: Wed, 14 Sep 2005 12:45:07 -0400 (EDT)
- To: "Ian Davis" <iand@internetalchemy.org>
- Cc: danbri@w3.org, "Ian Davis" <iand@internetalchemy.org>, "Christopher Welty" <welty@us.ibm.com>, "Lars Marius Garshol" <larsga@ontopia.net>, public-swbp-wg@w3.org, public-swbp-wg-request@w3.org
> danbri@w3.org wrote: >> That would be my preference, but I'm not sure the decision should be >> taken on a vocab-by-vocab basis. OWL Full systems I suspect are more >> likely to be comfortable with Web crawling. I expect it wouldn't be >> easy to get consensus on this, as the decision could be seen as >> answering the question "which is the default/main way of using OWL, >> and which is the niche interest specialist version? ...". > > Fair point. > > I guess it's too much to ask for DL systems to throw away triples that > don't make sense - probably can't be done at the individual triple level. I'd prefer them to ignore (and maybe not pass on) what they don't understand, and have considerable sympathy with the view Pat Hayes expresses in http://www.aifb.uni-karlsruhe.de/~sst/is/WebOntologyLanguage/hayes.htm That said there's probably a good reason why it can't be that easy (remembering huge 'dark triples' threads between RDFCore and WebOnt WGs re SW layering...). > Is this a candidate for best practice work - i.e. what schemas to put > where and how to select what you want. Yes and No. Yes - this is very much a practical issue standing in the way of SW deployment No - we can't really claim to be documenting a "best practice" before there's consensus in the community about what that best practice is :) > Just occurs to me that we're talking about different representations of > the same resource, which suggests that content negotiation may have a part > to play. That would really need different media types for the different > OWL flavours. I've wondered about that, but usually in a "yuck" context ;) It could be a reasonable design though I fear for how it might interact with semantics of fragment IDs that mean something in only one of the variant forms, eg. http://example.com/myvocab#foo where #foo is only defined in the Full representation of the resource. Maybe work-roundable but would need careful documentation. cheers, Dan > Ian > >
Received on Wednesday, 14 September 2005 16:46:19 UTC