Re: #rdfms-identity-anon-resources

Brian McBride wrote:
> 
> Dan Connolly wrote:
> [...]
> >
> > p.s. This message is really about an existing issue
> > http://www.w3.org/2000/03/rdf-tracking/#rdfms-identity-anon-resources
> 
> This is a key issue for getting the abstract syntax nailed down.  Since we
> are talking about it anyway, and its timely, lets activate it.  DanC, would
> you like to own it?

Well, I would, but I gather that means proposing a resolution
including test cases, and I don't see how to represent this
issue with test cases of the sort we've been doing;
this isn't a question of what n-triples you get from an RDF
document; it's a question of what the n-triples (specifically
the _:xyz terms) mean.

My position is represented by my n-tripes2kif.pl code; i.e.

	_:x <...prop> "val".
means the same thing as
	(exists (?x) (...prop ?x "val"))

i.e. to resolve this issue is to decide the formal semantics
of RDF. I guess I can work with Graham on this abstract
syntax and formal semantics stuff. [Graham, please let's
keep reification outa there.]

btw... in consideration of owning this issue, I reviewed
the discussion linked from the issues list, and the
case against digest/skolem URIs is made pretty well in:

  Arguments against digest URIs
  From: Jonas Liljegren (jonas@paranormal.o.se)
  Date: Sun, Jan 02 2000
  http://lists.w3.org/Archives/Public/www-rdf-interest/2000Jan/0001.html

in sum: we could, perhaps, encode all the information
about an existentially quantified term in some
sort of hash, but it would have to be a hash of the
whole formula/document, not just of the triple(s)
in which the term appears. And to do that is
at least very costly and constraining on implementations,
if not paradoxical/impossible.


>  (If know its hard for you to make it
> to the teleconferences - but I think we can work round that)


-- 
Dan Connolly, W3C http://www.w3.org/People/Connolly/

Received on Wednesday, 27 June 2001 02:14:35 UTC