Do we need a new attribute?

Hi,

some facts from RDF Core:

- anonymous resources are part of the RDF Model (they are part 
   of the MT and of N-Triples)
- anonymous resources are 'secure' in the sense that they are 
  not identified with other resources while merging RDF graps (see MT)

However, not all valid RDF graphs (accoring to MT or N-Triples) can
be represented in XML (see Sandro's last mail [1] or [2]). Two 
possibilities:

1. do nothing ;-)
2. enforce equivalence between MT,N-Triples and XML Serialization.

One possibility to achieve 2. is to extend the grammar with an 
attribute rdf:IID (InternalID) that must be interpreted by a 
parser as an anonymous resource. (That means, this ID is by definition
*not* stable and the parser may (and will) calculate a new internal ID
of it). RDF serializations according to the current spec would still be 
valid, but current parsers would have to be modified.

Another possibility to achieve 2. is to revise the MT and N-Triples
specs and to define: A valid RDF graph is an RDF Graph (according to
MT) with an existing XML serilization. (However, this is just a
definition,
it won't help in real live).

Best regards,
Stefan


[1]
http://lists.w3.org/Archives/Public/www-rdf-comments/2001JulSep/0302.html
[2] http://lists.w3.org/Archives/Public/www-rdf-logic/2001Sep/0000.html

Received on Thursday, 27 September 2001 04:13:23 UTC