W3C home > Mailing lists > Public > w3c-rdfcore-wg@w3.org > May 2002

RDF Graph (new version)

From: Jeremy Carroll <jjc@hplb.hpl.hp.com>
Date: Wed, 29 May 2002 09:52:04 +0100
To: "patrick hayes" <phayes@ai.uwf.edu>
Cc: <w3c-rdfcore-wg@w3.org>
Message-ID: <JAEBJCLMIFLKLOJGMELDMEANCEAA.jjc@hplb.hpl.hp.com>



Basically I am now out of time on this one. I will be able to spend an
hour or two on it while in Sardinia a week after next but I have no more
time this week or next.

Where I have got to is:

- a draft I am happy with except for:
  - the links to the test cases which are in the right place in the
document but are not are correct links (i.e. they link to the wrong
tests)
  - the ref to XML 1.1 that I've not done
  - there is a missing ref to a standard graph theory text.
    The one I use (Stewart?) doesn't really treat directed graphs, so I
would prefer to refer to a different one.


There are one or two stylistic issues, Pat touched on one:

> Quick comment: its dangerous to simply say that the RDF graph *is* a
> graph, even in brief, as in 2 para 1.

As a style I have started each part of the description with a brief less
formal statement (e.g. "A string literal is a string paired with a
language tag.") This is intended as a useful oversimplification expanded
upon in the ensuing text). Similarly the text that Pat picks up on is
clarified ad nauseum (in my view).

Perhaps that particularly piece could be improved by linking the words
RDF Graph to the fuller definition in section 2.7.

I think the most important stylistic question is that raised by me at
the end of the text about "a system of typed objects with identity"

Another issue that we have not discussed is the normalization
requirement on XML Literals. I arbitrarily chose a Note "If
compatibility with XML version 1.1 is desired ..." which fits with our
treatment of the composing character at the beginning of string
literals.

I hope there is enough there to have a discussion at this week's
telecon. The issues I think we should check we have consensus on are
basically those already highlighted:

- no isolated nodes
- any URI can be a property (although note the RDF closure rule in the
model theory that allows one to infer that the URI is of rdf:type
rdf:Property)
- XML Literals have no normative (unicode) normalization requirement,
but we have an interoperability note.

If the chair believes it would be worth formalling recording such
decisions I would be prepared to formally propose a motion on each.

Other than that I think there may be editorial issues with my choice of
words but I believe that the substantive content reflects working group
decisions.

Jeremy


-


Received on Wednesday, 29 May 2002 04:52:54 EDT

This archive was generated by hypermail pre-2.1.9 : Wednesday, 3 September 2003 09:48:16 EDT