From: pat hayes (phayes@ai.uwf.edu)
Date: 07/09/01
>pat hayes wrote: > > > I would like to see the 'URI' point retained in the document. The > > more the coreWG gets its node rubbed in this issue, the better :-) > >The URI point from Peter's notes stated: > > > - what is a URI? - syntax and semantics > >URI syntax is specified in http://www.ietf.org/rfc/rfc2396.txt by Tim et al. >so I don't really see the substance of that part of the complaint. >(The RFC describes the syntax that is common to all URI schemes, >and the scheme-specific details of the syntax are/should be specified >in other places, seems reasonable to me). > >Concering semantics, the same document says: >"A Uniform Resource Identifier (URI) is a compact string of characters > for identifying an abstract or physical resource." > >I've always understood that URI's are simply globally unique names, Right. But this isnt yet sharply enough defined. Do you mean 'logical names' (ie constant symbols), or something more like 'names' in NL? The discussions surrounding the notion of URI often seem to blur the two. Also , http://www.ietf.org/rfc/rfc2396.txt consistently says that URI's are a "means for identifying" a resource, and this innocent-seeming phrase hides a huge array of assumptions. File names for example provide a means to identify their files in a way that personal names do not identify the people they name, and personal names (and public names more generally) identify people (and cities, etc.) in ways that logical 'names' do not, and probably cannot possibly, emulate. Merely denoting something is not like being able to pick it out from a police line-up. So my conclusion is that this apparently technical area is in fact a mass of confusion and unstated assumptions, some of which are widely understood (but not widely agreed on) and others not properly discussed anywhere. Finally, even if we take the very restrictive view and say that URI's are just logical constants with an odd syntax (required to avoid global name-clashes), then the question arises as to what the nonlogical vocabulary of a language like RDF or DAML+OIL is supposed to be? Validity has to be understood relative to a particular vocabulary. We have had some very confused exchanges on RDFCore recently about whether RDF 'anonymous' nodes are like existential varaibles, for example, which have gotten confused partly because folk do not realize that skolemisation is not valid because it changes the logical signature. >(which I think is pretty much what Tim's RFC says). In terms of our >model-theoretic semantics, are URI's not simply the elements of AD, >and that's it? Certainly not, in most interperations. (One in which the elements of AD were URI's would be the DAML equivalent of a Herbrand interpretation.) >So I felt unsure as to what really the substance of Peter's point was. >I also reread much of the discussion on URI's on rdf-logic, back in April >(starting from >http://lists.w3.org/Archives/Public/www-rdf-logic/2001Apr/0261.html) >but I found the whole discussion rather unilluminating (then, and >now again), mostly by folks suffering from a severe case of >use\mention confusion. Right, and that is another issue I would like to rub the WG's node in. >Pat, can you expand on what exactly the coreWG should get its nose >rubbed into? See above. Pat --------------------------------------------------------------------- IHMC (850)434 8903 home 40 South Alcaniz St. (850)202 4416 office Pensacola, FL 32501 (850)202 4440 fax phayes@ai.uwf.edu http://www.coginst.uwf.edu/~phayes
This archive was generated by hypermail 2.1.4 : 04/02/02 EST