- From: Sergey Melnik <melnik@db.stanford.edu>
- Date: Fri, 15 Jun 2001 18:58:11 -0700
- To: Dan Brickley <danbri@w3.org>
- CC: Aaron Swartz <aswartz@upclink.com>, jos.deroo.jd@belgium.agfa.com, Jan.Grant@bristol.ac.uk, w3c-rdfcore-wg@w3.org
Dan Brickley wrote: > > On Fri, 15 Jun 2001, Aaron Swartz wrote: > > > On Friday, June 15, 2001, at 06:10 PM, Sergey Melnik wrote: > > > > > I don't agree that anonymous nodes should be part of the abstract > > > syntax, and would suggest to consider this issue when cleaning up the > > > model. > > I disagree: it is critically important to distinguish between well known, > public URI names for things and ad-hoc generated placeholder IDs that have > been dreamt up by an RDF/XML parser. Unless the abstract syntax (or > whatever we call it) maintains that distinction, we risk getting into a > terrible muddle. > > > I tend to agree with this position. However, I would take it one > > step further -- I believe that these "uniquely generated > > resources" should have consistent, repeatably generated URIs. > > That is, all parsers should assign the same genid to the same > > resource. > > do you really mean this last claim? > I suspect you meant that all parsers should assign a predictable genid > given a common RDF/XML description mentioning a resource. 'all parsers > should assign the same genid to the same resource' would be magic, since > many times parsers won't have that information accessible. Dan, just to provide a pointer, a whole while ago I proposed an algorithm for doing exacty that "magic". A summary of the proposal can be found at http://nestroy.wi-inf.uni-essen.de/rdf/sum_rdf_api/ under "URI generation for anonymous resources" (by Peter Hannappel and Reinhold Klapsing). I believe, if needed, this algorithm can be tweaked to eliminated the problems wrt using XOR (as pointed out by Brian long ago). I still think anonymous resources first need to be resolved as a model issue. Then, we can move to the syntax. BTW, well-known public names introduce another danger, which can be even more malicious that generating ad-hoc URIs: if I'm using danbri@w3.org as an identifier for Dan's email address, and (mis)using for identifying Dan himself, I can get into a worse trouble as compared with using two ad-hoc, not as well-known, but sufficiently unique URIs that allow me to distinguish that I'm dealing with two separate entities. When I have two different resources, they still may turn out to represent the same thing. In contrast, if I'm using a single resource for representing several different things, I'm in a bigger mess. Sergey
Received on Friday, 15 June 2001 21:32:03 UTC