Re: profiles/modules/levels

What I'm trying to ensure is that SpecGL doesn't stifle innovation. What I
meant about definitions is that I don't want to give people the impression
the our prescribed p/m/l are the only correct ways to implement DoV. It
was mentioned yesterday that p/m/l are commonly used DoV concepts and so
we should work hard to define them carefully in SpecGL. Then, if people
want to use them they will have a consistent definition (as you argue) -
and I agree with that.
But what if the spec author choses another type of architecture that we
have not thought of? My argument is about this case. I want to make sure that our
checkpoints address this possibility. At the moment we have specific
checkpoints for p, m and l. But if someone wants to use a new type of Dov
called 'personality' or whatever then SpecGL is silent. So I am arguing
that when we roll the checkpoints into a single GL, we should keep the
specific checkpoints for the important concepts of p/m/l but ensure that
we have general checkpoints too.

-Andrew

On Mon, 28 Apr 2003 david_marston@us.ibm.com wrote:

>
> Andrew Thackrah writes:
> >...since it seems from QAWG discussion that there are differences of
> >opinion on our definition of p/m/l then I don't think we are in a
> >position to impose a rigid definition on others....It doesn't matter if
> >your DoV does not conform to someone elses definition of a profile or
> >whatever - all you have to do is document your chosen system and if you
> >have more than one DoV then document the relationship between them.
>
> But it *does* matter, becausae these specs aren't written in isolation,
> but (usually) to be part of an integrated Web system. Schema Part 2
> defines data types, then XPath builds expressions around those types,
> then XForms and XSLT use XPath expressions, etc. If some data types
> (e.g., the whole ID-IDREF bundle) are designated as an optional
> module, then specs building above that need to say whether they depend
> on the full set of types or just the "core" set. QAWG has also talked
> about how profiles can be assembled from modules, so the naming of
> subsets is useful even within a single spec.
>
> Some WGs may have used the p/m/l terminology in ways other than the
> SpecGL-sanctioned meanings in the past, but it's desirable that the
> W3C move toward consistent usage. Notice that the documentation terms
> "Version", "Edition", and "Part" have been subject to consistency
> constraints for some time now, and the specs are better for it.
> .................David Marston
>

Received on Tuesday, 29 April 2003 09:33:35 UTC