W3C home > Mailing lists > Public > public-rdf-dawg@w3.org > January to March 2004

RE: thoughts and some refs about AFS-2 UC (simplicity, minimalism)

From: Dan Connolly <connolly@w3.org>
Date: Fri, 19 Mar 2004 09:16:40 -0600
To: "Seaborne, Andy" <andy.seaborne@hp.com>
Cc: RDF Data Access Working Group <public-rdf-dawg@w3.org>
Message-Id: <1079709400.2471.63.camel@dirk>

On Fri, 2004-03-19 at 08:37, Seaborne, Andy wrote:
> -------- Original Message --------
> > From: Alberto Reggiori <>
> > Date: 19 March 2004 08:58
> > provenance/attributions seem to be a key aspect to consider for the
> > success of the whole RDF and associated query languages stuff - anybody
> > else agree on this?
> I agree that in the overall swemantci web architecture
> provenance/attributions are going to be important and that it is very
> significant for query in the long run.  I don't think that means we/DAWG
> must address it in order to get a rec done that helps many people and
> systems.  I think that is huge benefit in a "simple" rec that helps toolkit
> builderd provide the same capabilities (they can confidently invest time in
> building a quality implementation); it helps application writers be able to
> access different sources without needing to adapt their applications to the
> details of a particular source, and it helps end users by allowing them to
> use their apps with many data sources.

Yeah verily!

A number of us said, in our introductions to the WG, that we'd like to
get something simple done soon. Let's follow that up with use cases
that focus on the very simple conjuctive query that we all *know*
deserves W3C Recommendation status ASAP.

My not-so-secret goal is Candidate Rec in six months. To do
that, we need to stick to functionality that we've already coded.
Twice. i.e. one or more of us has done the prototype, thrown it
away, and built it for real.

We have time to figure out what to do with the arbitrary
differences between the implementations we've got (i.e. one
implementation uses a comma where another used a period; one
supports integer add and one does float) and
to spell-check our spec, set up a test harness, and that's
about it. In the CR phase, we can flesh out the test suite
and evaluate whether this minimal functionality is too minimal
to be useful; i.e. whether that set of features should go
to REC as version one, or whether we need to add more features
before advancing.

Our charter calls for Proposed Recommendation in June 2005.
Let's aim for that to be Proposed Recommendation of version 2, with
some of the more interesting features... let's polish
our provenance design while we're in CR on the first version,
i.e. while the marketplace is solidifying.

In the WebOnt WG, we distinguished requirements from objectives.
Requirements were just that: the WG agreed that the design
isn't done unless it meets the requirements. Objectives were
things that everybody agreed would be nice, but we weren't
guaranteeing we'd get them done in this version.

Provenance doesn't look like a requirement to me.
Of course I think it's interesting and key to the future.
I spend a lot of time researching it and doing advanced
development with it. But it doesn't looke like part
of the so-called "minimum required to declare victory."

That might look like an opinion that goes beyond the chair's
neutral role. In some ways, it is: the chair also has some
technical leadership rights and responsibilities. But in other
ways, I'm just advising the group on how to get what you said
you want in your introductions.

Dan Connolly, W3C http://www.w3.org/People/Connolly/
see you at the WWW2004 in NY 17-22 May?
Received on Friday, 19 March 2004 10:16:41 UTC

This archive was generated by hypermail 2.4.0 : Friday, 17 January 2020 17:00:24 UTC