toward 3 WDs this month (MT, RDF/XML, test suite) [was: RDFCore telecon agenda for 2001-08-17]

Brian McBride wrote:
[...]
> Discuss the strawman document structure: (15 mins)
> 
>   http://lists.w3.org/Archives/Public/w3c-rdfcore-wg/2001Aug/0038.html
> 
> Goal - identify issues and concerns.  So far we have:
> 
>   o we need to get started and see the material before getting into
>     lengthy theoretical discussions
> 
>   o how much 'explanatory' material should the spec have
> 
>   o which are we revising/clarifying and which are new work products
> 
>   o timescales/resources
> 
>   o don't make folks read two documents
> 
>   o do we need an executive overview
> 
>   o where do n-triples fit in the spec
> 
>   o moving containers/reification may cause problems with review process
>     (what problems?)
> 
> Propose we take a shot at a first draft of a core model spec based on
> the current M&S document merged with the appropriate sections of the
> model theory.

Please, no! As a reviewer, I'll find it very hard not to glaze
over the old text while looking at a merge with the model theory.

Let's release the model theory as a working draft on its own,
with something in the Status about how were considering merging
it with the RDF 1.0 spec or whatever.

We're well overdue to publish *something* (W3C process calls
for a publication at least every 3 months; this WG was
chartered 30 March and we haven't published yet. The Director
should be knocking on our door any day now wondering what's
up).

I'm frankly not in the mood to take on the huge task of
rewriting (or even reviewing a rewrite) of the RDF 1.0 spec
any time soon.

We're obliged to produce errata. We're doing that (sorta;
we need to move the relevant stuff from the issues list to
the official errata page at some point... soon, please?)

Errata aren't very exciting/useful to developers, but test
cases are. I think the way to best share the work we've with
the developer community is to make noise around a release
of the test cases we've assembled so far. Art, Brian, would you
be interested to write a 1 or 2 page covering document
for the test cases, discussing their relationship to
the errata and issues list, to be released as a WD? Anybody?

And let's release the model theory to make the logic geeks
happy; it does address many of the more thorny issues on
our issues list.

And let's release the work we've done on the XML serializtion
of RDF as its own WD. (where is that thing? I don't
see it linked form http://www.w3.org/2001/sw/RDFCore/ ...
ah... there it is...
  RDF/XML Syntax Grammar Experiments Dave Beckett (Fri, Jul 20 2001)
  http://lists.w3.org/Archives/Public/w3c-rdfcore-wg/2001Jul/0276.html
)

Let's do that much this month. Three WDs by the end of Aug
would feel good to me.

After that (unless somebody produces something sooner),
let's do something for the end users: a tutorial, a
collection of success-stories/case-studies, something. Or
maybe organize an RDF users workshop or something.

Then let's revisit the RDF 1.0 spec; by then, we should be
able to cut 60% of the prose: the introductory/motivational
stuff, etc. Or maybe it gets completely superceded by
the model theory and the RDF-in-XML grammar thing.
(ok, well, we'll have to figure out where reification and
collections go. My vote is: a separate spec for each.)

I haven't given much thought to where RDF schema fits in
this schedule. Salt to taste.

-- 
Dan Connolly, W3C http://www.w3.org/People/Connolly/

Received on Thursday, 16 August 2001 13:00:02 UTC