Cart
[Home] [By Thread] [By Date] [Recent Entries]
If I misstate the following, anyone feel free to correct me. I agree with you, Uche, but some background follows. Simon says: >> Bill de Hora posted something on www-tag that's very much worth >> contemplating, even if you don't like it. >> >> http://lists.w3.org/Archives/Public/www-tag/2003Jan/0301.html From: Uche Ogbuji [mailto:uche.ogbuji@f...] >I'm having a great deal of trouble discerning anything new here. This seems >to me jsat a different stating of my own frequently-stated belief that >semantic laxness rather than rigor is essential to make URIs work. <snip /> >I've always said the sky is not falling in my corner of the URI universe, and >precisely because I've pretty much always admitted the many-to-many idea, and >scoffed at the very concept of a truly universal identifier. Because it is an issue that resurfaces again and again with respect to the definition of what a "resource" is. Apparently in the RDF formal definitions, it doesn't work because RDF or KR needs a stricter one to one mapping and URIs don't formally provide that for tangible objects: just resources. I agree with what you are saying but there exists definitional or formal confusion. The TAG has been doing another round on this one. It comes down to unique identification of a "resource" where a resource is "anything with identity" but resource is a term for a concept, and that concept braces representations. At the end, the URI is naming a concept, not an object. This works because the implementors understand that and where they see "resource" they use, as Joshua Allen says, a "hypertext dispenser" or one might say, a representation dispenser. The question becomes, who should fix what? Empirically, the URIs on the web work and implementors do the right thing. In RDF, that does not seem to sufficient and then the argument becomes, what if anything should the RDF spec team fix in their specification? I've no skinny in this game, but the arguments are very intriguing to follow. len
|
