some comments on the new RDF document

I have a few comments on the new RDF document, and specifically on 
suggestions made about interactions with the Primer.

*  I think the title of the new document, "Resource Description 
Framework", is misleading (this is the second time I've said this).  ALL 
our documents are about the "Resource Description Framework".  Can we 
please think of a more descriptive (and specific) title (assuming we can 
agree on what the document is supposed to do)?  (Unless I'm mistaken, I 
think Brian has suggested "RDF Data Model" for this.  I don't 
particularly care for this title either!)

*  Brian seems to suggest that sections 2.1 (motivation) and 2.2 (design 
goals) of the new document should go in the Primer.  I can see how that 
makes sense, but I'll have to think about whether those sections could 
be just plunked into the front of the Primer more or less as is (would 
this include all of section 2.2's subsections?)  Opinions (I don't feel 
strongly one way or the other)?

*  Graham has suggested that section 2.3.1 of the new document (on the 
Model Theory) should stay where it is, and the corresponding material 
might be removed from the Primer (Section 7.1).  This and section 7.2 
(describing the Test Cases document) were originally in the Primer with 
the idea that the Primer was going to have to provide a small bit of 
introduction to each of the other documents in the RDF "suite" (say why 
the document is there).  With the understanding that the new document 
was going to serve that purpose, I suggested to Graham that both Primer 
sections 7.1 and 7.2 could go in the new document.  On the other hand, I 
notice that Brian seems to think the new document should NOT describe 
the other documents (although I may be misinterpreting his comments on 
this subject).  The WG may want to express an opinion about this;  I 
don't particularly care one way or the other.  I do think that some 
document, either the new one or the Primer, should briefly introduce the 
entire suite.

*  Section 4.4 of the new document is titled "RDF triples".  This and 
section 4.5 seem to take a somewhat different line to describing the 
graph model than the other documents (but I may be wrong, and this 
approach may be better), in that it introduces triples first, and then 
describes the graph as being a collection of triples, rather than 
describing the graph first, and describing triples as a way of encoding 
or talking about the graph.  The former seems to make triples seem 
somehow more fundamental than the latter.  On the other hand, talking 
about triples first does make it easy to see why you can't have a graph 
just consisting of a node.

*  Graham has suggested that section 5.1.1 of the new document 
(containers vs. multiple properties) should go in the Primer.  I 
deliberately avoided putting this material (from the M&S) in the Primer 
when I wrote the container section (and said so in a comment), because I 
thought its explanation was mostly dubious.  Looking at it again, I can 
see putting it in the Primer, but I would want to rewrite it (not a lot 
of work, IMO).  Opinions?

*  Graham has suggested that the Primer should NOT cover reification, 
and that section 5.2 (on reification) in the new document belongs in the 
vocabulary/schema document.  I'd certainly be happy not to cover 
reification in the Primer.  Opinions?

*  Graham has suggested that the Primer not cover rdf:value, and that 
section 5.5 in the new document can be dropped, as being covered in the 
vocabulary/schema document.  I don't mind not covering this in the 
Primer, but at one time we had some explicit direction from the WG to 
cover it.  Has that been changed?  How important do people think talking 
about this in the Primer is?

*  Graham notes that section 6.2 of the new document (RDF in HTML) could 
go in the Primer.  I agree.

*  Jeremy notes that section 6.3 of the new document (Fragment 
identifiers) has not been discussed within the WG.  I'm going to want to 
include some such material (once the WG looks at it) in the Primer, 
whether it remains in the new document or not.

*  Graham notes that sections 6.4 (URI references from Qnames) and 6.5 
(Boolean valued properties) of the new document might be better in the 
Primer.  I agree about URI references from Qnames.  About Boolean valued 
properties:  what is the justification for saying anything explicit 
about this at all?  There are lots of other values RDF doesn't define 
either!

--Frank

 


-- 
Frank Manola                   The MITRE Corporation
202 Burlington Road, MS A345   Bedford, MA 01730-1420
mailto:fmanola@mitre.org       voice: 781-271-8147   FAX: 781-271-875

Received on Wednesday, 3 July 2002 15:46:44 UTC