Re: Issue #22, #12, and horizontal groups

I use the term "substantive/vertical WGs" to mean those WGs that are
working toward a Recommendation of a technology for implementation by
non-W3C developers, like XSLT, SVG, etc.

Lofton quoted me and responded thusly:
>At 12:58 PM 1/7/02 -0500, you wrote:
>>[...]
>>Issue 12 (raised by me) is W3C-internal; the question is about QAWG
>>interacting with WAI and I18N. In my mind, the question is also
>>whether the W3C as a whole begins to think of "horizontal groups" as
>>a genre before they really should.
>
>Rationale for the latter?  (E.g., so as not to preclude other
>organizational solutions?)

The product of a substantive/vertical WG must address various global
concerns. Some of those needs are addressed by having a candidate
document reviewed by a separate WG or TAG or whatever. Does the W3C
intend to have a meta-process that each new concern will be
represented by the creation of a horizontal WG? If so, should QAWG
be fit into that mold, meaning QA is yet another step alongside
WAI, I18N, etc. reviews? Or if not, would the meta-process (or some
body such as the AB) prescribe alternate ways to see that the
global concern influences the work of every substantive/vertical WG?

The original issue 12 really asks about whether the QA concerns also
affect how WAI and I18N do their horizontal reviews. Does QAWG want
to give tools/methods to the horizontal WGs as well as the vertical?
Does QAWG want to be outside the whole horizontal/vertical paradigm?

We have talked about how concern for testability and conformance
measurement should pervade the work of a substantive/vertical WG
right from the start. All QA professionals, in software and other
fields, know the dangers of thinking that quality is "added on"
after the foundation design is done. If the first documents from
QAWG imply that it should be yet another horizontal group, that
posture would invite the substantive/vertical WGs to externalize
their quality concerns. The QAWG might actually get to review each
Candidate Recommendation, but that potential should be presented
as one of several under consideration. The dominant signal from
the FPWD to each substantive/vertical WG should be "you will need
to think about testability in everything you do from now on."
.................David Marston

Received on Wednesday, 9 January 2002 14:27:40 UTC