Comments for RDF series Last Call Working Drafts

Here are comments on your RDF Last Call series [1,2,3,4,5,6], to use or
not as you see fit. I am sending notes on each part separately and hope
they get to you in time; I am still reading RDF/XML. This note applies
to all parts.

I would prefer to see one RDF Recommendation. It could be updated in
"editions," and its components could be new "versions." Some random
dependencies that I ran into in a quick reading: Schema semantics are 
in semantics. Semantics 0.2 says semantics uses
N-Triples which are in Tests. Concepts reproduces an example from the
Primer. Everybody depends on Concepts. Concepts "3.7 RDF Core URI
Vocabulary and Namespaces (Normative)" points to Syntax and Schema for
the vocabulary. Schema duplicates some of Semantics. The Primer
duplicates some of Syntax. The Primer, Semantics, and Syntax all depend
on the N-Triples language definition in Test Cases. Schema depends on
Syntax. With things so intertwined, one spec with chapters and a
single-page alternate format would be a relief.

External links within the prose need links to references for people who
print: http://www.w3.org/2001/06/manual/#linking-within. If this was one
spec you would have far fewer external links, and that would make some
parts more readable.

Question: Is "URIref" W3C or RDF Core Working Group jargon or is it
accepted universally? I can see it from Google in one of TimBL's slide
sets and in one reference each at DAML and OASIS. It is a stumbling
block to a newcomer to have to learn it unless necessary. "URIref" is
in the Primer over a hundred times. It is all lowercase "uriref" fifty
times in Semantics which says it is defined in Concepts. But Concepts
mentions "URIref" twice, once in bold blue, in non-normative sections.
So I don't see evidence that it is either defined or needed.
P.S. I just found URIref defined in the N-Triples BNF in Test Cases
which depends on RFC 2396 which says "URI references." At least,
Semantics could get its definition from N-Triples instead of from
Concepts, or Concepts could define URIref. The Primer, Semantics and
Concepts could agree on its spelling.

Semantics, the Primer, Concepts and Test Cases would be 10% smaller run
through Tidy with indent off (Syntax and Schema already do something
like this).

The dark red link color in Semantics could be confused with the
definition color in Syntax. One CSS style sheet for the whole series
might help, or maybe better, you could just use the required W3C TR
style sheets with an addition for the example divs.

All of the tables could have cellpadding="4" or so. Some do already.

Name in References [RDF-VOCABULARY] should be [RDFS] or [RDF-SCHEMA],
or else you could refer to the spec as "RDF Vocabulary."

None of RDF has conformance sections and it may be that they all should
have them. (Syntax mentions RFC 2119, conforms to Infoset otherwise has
none. Concepts, Schema none. Semantics mentions RFC 2119. Tests
mentions that N-Triples is used for conformance.)

Schema says super-class. Semantics says superclass. I'd pick one.

If you do issue RDF in parts, you'll need a uniform section in each
document that links to the other parts. Some do already. This should be
the same text, in the same order, in the same place, at the beginning
in each document. If you want to call RDF Schema "RDFS," establish that
here.

For Acknowledgments: participants in a Working Group are termed
participants not members to avoid confusion with W3C "Members." I
believe Agfa is all lowercase except in Poland, and Hewlett-Packard is
always hyphenated. De Roo is spelled about three different ways (DeRoo,
deRoo, de Roo). I think De Roo is correct.

[1] http://www.w3.org/TR/2003/WD-rdf-primer-20030123/
[2] http://www.w3.org/TR/2003/WD-rdf-concepts-20030123/
[3] http://www.w3.org/TR/2003/WD-rdf-syntax-grammar-20030123/
[4] http://www.w3.org/TR/2003/WD-rdf-mt-20030123/
[5] http://www.w3.org/TR/2003/WD-rdf-schema-20030123/
[6] http://www.w3.org/TR/2003/WD-rdf-testcases-20030123/

Best wishes for your project,
-- 
Susan Lesch           http://www.w3.org/People/Lesch/
mailto:lesch@w3.org               tel:+1.858.483.4819
World Wide Web Consortium (W3C)    http://www.w3.org/

Received on Monday, 24 February 2003 00:50:35 UTC