Re: Options for dealing with IDs (RDF)

On Thursday, January 9, 2003, 4:22:23 PM, Sandro wrote:

>> So the RDF "ID" attribute was never an XML ID.

SH> I think this was a mistake in the design of RDF.

Hi Sandro.

SH> (Of course there was and still is no way to tell an XML processor that
SH> it's an id [that's the main subject of this thread], but it should
SH> have been made clear that conceptually it was an XML id.  People can
SH> and do make DTDs for limited vocabulary subsets of RDF/XML, and those
SH> DTDs should [in a more logical universe] indicate that rdf:ID is an
SH> XML id.)

I agree that there was no way and that making such a way is the point
of this thread.

SH> The RDF/XML ID should [IMHO] be used like other fragment identifiers
SH> to point to a place in or section of some document instance (as in
SH> HTML, XML, VRML, sound recordings, etc).  If RDF wants to talk about
SH> the subject of the description there, it should do so explicitely.

I tend to agree but thought I would get burned at the stake for saying
so. Since you have boldly stated it then I will (brave brave sir
robin) step up second and agree.

SH> For more details on how to do this with very little pain, see my
SH> Disambiguating RDF Identifiers [1].

SH> To tie this into the other two recent www-tag threads, this problem
SH> really glares in content negotation and namespace documents.

YSH>  What am
SH> I going to fetch from some namespace address?  Maybe HTML (as in my
SH> RDDL proposal [2]), but I'd also like to content-negotate getting the
SH> same information in RDF/XML.  But with the currently proposed
SH> application/rdf+xml media type, I can't do that (as the TAG recognized
SH> [3]).  So the web architecture is broken here....

Ah, interesting point. I still think that fragids and content
negotiation is a broken hole, but agree that in this particular case
that would heal an important part of the hole (while still leaving it
open in the general case).


-- 
 Chris                            mailto:chris@w3.org
 Since I meant to add this on my previous RDF/XML post but forgot
 "So if it weighs the same as a duck, then it must be made of wood,
 and therefore ... its an RDF:ID!