Re: refactoring RDF/XML Syntax

Hi Shelley,

Thank you for your question.

Shelley Powers wrote:
> 
> I understand the interest in improving the RDF syntax, but I'm confused as to the
> process the working group is taking.

The WG's goals and objectives are best understood by reading our charter:

  http://www.w3.org/2001/sw/RDFCoreWGCharter

where it says:

  The RDF Core WG is neither chartered to develop a new RDF syntax,
  nor to reformulate the RDF model.

What I want to make clear is that we are *not* improving the syntax itself,
we are improving the *specifiction* of the syntax.

> 
> Wouldn't the process of updating or modifying a released specification be to start a
> new release version of the specification? In other words, since the RDF Model and
> Syntax are currently a W3C initial recommendation, wouldn't new effort that actually
> makes modification to this specification be release 1.1, or even perhaps 2.0? This
> isn't errata -- this effort is a redesign of parts of the specification.

I expect us to release new specifications.  The new syntax WD is a step on
the way to producing a new RDF/XML syntax recommendation.  My hopes for that
recommendation are that it will focus on describing the syntax only; the
description of the RDF model will be clearly separated, and that it will 
achieve two objectives:

  o it will describe both the RDF/XML language and its translation into triples
    more precisely than the current specification

  o it will be easier to understand than the current specification

One of the key things in the new WD is the specification of RDF in terms of 
infoset, rather than in terms of characters.  We believe this is a signifcant
step forward in formalizing the description of the langauge.

> 
> I can't help thinking that incorporating these changes into a new release of the
> specification provides some stability in your process, as well as decreases
> confusion; folks will reference a specific release when discussing a certain feature
> of the RDF syntax.

That is what I expect us to do.  From the charter, our deliverables include:

  * update the RDF Model and Syntax Specification (as one, 
    two or more documents) clarifying the model and fixing
    issues with syntax

That said, producing a new specification takes time, and where we can resolve
issues that folks have raised and are recorded in the issues document:

  http://www.w3.org/2000/03/rdf-tracking/

we will continue to do so.  Those resolutions will be incorporated into the
new specifications.  Such decisions, generally, are clarifications of
the current specification.  For example, since the grammar was ambiguous,
it was not clear whether an rdf:about attribute was legal on a container.
We have resolved that ambiguity.

You should be aware, however, that those decisions may change before the
final specifications are produced.

> 
> I am in the process of using RDF for a couple of specific production issues, and
> modifications to an existing release make it difficult for me to sell the use of RDF
> to management. However, work on a new release is pretty standard for a W3C working
> group and doesn't generate quite the same concern.

We are working on new releases of the specifications.

It cannot be denied that a number of issues have been raised with the
current M&S specification.  These are documented in the RDF issues list:

  http://www.w3.org/2000/03/rdf-tracking/

It is for you and your management to assess the impact of the presence
of these issues on your own development plans.  I hope the fact that we
have a public list of them will help folks to assess whether they cause
a genuine problem for their application, and will help folks simply
avoid using some of the trickier bits of the spec until the WG has
resolved the problems.

I hope this response has been of some help to you.  Please come back if
this isn't what you need.  I would really like to understand what we
can do to help folks in your situation.

Brian McBride
RDFCore co-chair

ps: If you can say more about your application of RDF, I'd like to hear
about it.  The better the WG understands how RDF is being used, the 
better we can make the tradeoffs that we inevitably face.

Brian

Received on Thursday, 13 September 2001 03:24:22 UTC