xml-dev - Re: Anti-Ranti

[keshlam@us.ibm.com]

> > Curiously, the official W3C process expects all non-members to wait
> > till "Proposed Recommendation" or (new) "Candidate Recommendation"
> > to express opinions.  An "appropriate" moment is already defined.
>
> That's not what the W3C Process Document section on WG deliverable
> says:

Seconded. That's NOT the intent.

The working groups very much want feedback as early as possible. We don't
have any vested interest in inventing all the details ourselves, and we
_REALLY_ don't want to get all the way to Last Call/CR/PR only to discover
that we overlooked a critical use case -- or even a simple/obvious/easy
detail. Early input is very much appreciated.

I believe that every WG has a public mailing list for discussion of the
working drafts. Use it. Use it well enough and volunteer enough time to the
cause, and you might even find yourself invited to join the Interest
Group... which in some cases differs from WG membership mainly in not
having a vote.

Of course that doesn't mean all suggestions will be accepted. Heck, not all
suggestions from within a WG are accepted; it's a committee, and there are
honest disagreements about what solutions are best. And sometimes schedules
and available manpower limit what can be done. But the comments and
questions on the public list _are_ read and taken seriously.


I'm having trouble putting my hands on it now, but someone's webpage
includes an "Outsider's Guide to the W3C" which is a good intro to the
process. Here's my own brief attempt. Feedback is needed at every stage!

Public Working Drafts (WD) are released specifically with the intent of
requesting public comment. (There are nonpublic WD's in between these,
mostly for use within the working group, tracking the spec's evolution.)

Last Call (LC) is a public WD and a public call for final sanity-check of
the overall design.

Candidate Recommendation (CR), added recently, is an "implementation
period". When a spec goes from LC to CR, that's a statement that the WG
thinks the design is stable and wants folks to go off and start building
code and reporting their experience. The goal is to make sure that the spec
really is complete -- no ambiguities or missing data -- and can be
implemented in a reasonable manner. CR  is also a promise that no more
major changes will occur unless a problem is exposed -- so it's a good time
for you to start working on your prototypes in any case, even if you
weren't willing to track the evolution of the WD. Generally, getting out of
CR phase requires that several successful implementations have been
reported... or that the unimplemented sections be dropped from the spec.

Ppoposal (PR) is an official request for the W3C members to vote on whether
this spec should be accepted as a W3C Recommendation. This is the one stage
where non-Members really don't have a direct voice; if you aren't a Member,
you can't vote.

Finally, if the votes are in favor, the specification becomes a
Reccomendation (REC), which is as official as the W3C ever gets. At that
point, the spec is pretty much cast in concrete... but is often still
looking for input for the next "level", or for other tools based on this
one, so the process repeats from the beginning.




______________________________________
Joe Kesselman  / IBM Research



***************************************************************************
This is xml-dev, the mailing list for XML developers.
To unsubscribe, mailto:majordomo@xml.org&BODY=unsubscribe%20xml-dev
List archives are available at http://xml.org/archives/xml-dev/
***************************************************************************