Article:		RPV: Triples Made Plain
Subject:		RPV comments
Date:		2002-11-21 02:07:00
From:		Ian Young
I'm not sure we should be commenting on something that hasn't even had the obvious typos fixed yet. Nonetheless... RPV as currently described lacks a way of referencing blank nodes in the serialization. Some mechanism like RDF/XML's rdf:nodeID is needed. You give an example of an <R> element with both id and r attributes. I think this is a misunderstanding: it only makes sense for one or the other to be given. In fact, since the id attribute is equivalent to RDF/XML's rdf:ID and the r attribute to rdf:about, only the r attribute is required in RPV --there is no difference in semantics between <R id="name"> and <R r="#name"> More work is needed on typed literals, particularly XML literals. Section 4 should be dropped. It is not necessary to introduce new syntax to encode RDF statements that introduce a 'reification'. At present, this section is also incorrect in that it requires the statement being reified to be asserted as well as its reification.

Previous Message

Next Message

• RPV comments
  2002-11-21 14:31:18 Kendall Clark [Reply]
  
  I trust you've sent these comments to Tim Bray? I'm not sure I agree about the r and id attributes of the R element, but no matter.

  
  I think you may have misunderstood the point of the XML-Deviant column, which is precisely to comment upon stuff that's being actively worked on and developed. Note that I didn't tell anyone to go out and bet the farm on RPV, merely that some parts of the recent RDF donnybrook have been productive, unlike most of the rest of it.