- From: Michael Kifer <kifer@cs.sunysb.edu>
- Date: Fri, 01 Jul 2005 11:09:07 -0400
- To: Dave Reynolds <der@hplb.hpl.hp.com>
- Cc: www-rdf-rules@w3.org
> Michael Kifer wrote: > > > First, it is not too late to fix the mistakes in RDF. As far as I know, the > > implementations of N3 don't respect the existential semantics of blank > > nodes. And you kept saying in this thread that N3 is an RDF language. > > Don't go there. > > First, even if there was some reason to change the specification of RDF, it > should not be within the scope of any putative rules working group to do so. > > Second, many RDF implementations do try to respect bNode semantics. For > example by representing bNodes in stored triples by skolem constants and > bNodes in queries by query variables. They will *also* allow programmatic > access to the implementation objects that represent bNodes for the purposes > of manipulation and editing but that doesn't detract from the existence of > conformant APIs. Ha, but this is only an approximation -- this is my whole point! So, they don't implement the semantics anyway. > > I'm not sure why you are particularly singling out N3, and I am not a CWM > expert, but my understanding was that CWM *does* attempt to respect the > semantics of bNodes. See above. --michael
Received on Friday, 1 July 2005 15:12:45 UTC