QA WG Teleconference minutes, Sep 13 2004

QA Working Group Teleconference
Monday, 13 September 2004
--
Scribe: Dominique Hazaël-Massieux (W3C)
Chair: Lynne Rosenthal (NIST)

Attendees:
(PC) Patrick Curran (Sun Microsystems)
(DD) Dimitris Dimitriadis (Ontologicon)
(DH) Dominique Hazaël-Massieux (W3C)
(LR) Lynne Rosenthal (NIST - IG co-chair)
(AT) Andrew Thackrah (Open Group)
(DM) David Marston (IBM)

Regrets: 
(KD) Karl Dubost (W3C, WG co-chair)
(LH) Lofton Henderson (CGMO - WG co-chair)
(MS) Mark Skall (NIST)

Summary of New Action Items: 
LR, DH to contact informally but personally a few people to get feedback
on new version SpecGL by 2004-09-20
PC to have a look at the thread on www-qa regarding our test principles
applied to GEO Tests, by 2004-09-20


New issues:
* our GP and Principles should have assessable results
* the principles and GL of section "Quality Control" in SpecGL needs to
reworded or downgraded

Agenda: http://lists.w3.org/Archives/Public/www-qa-wg/2004Sep/0006.html
Previous Telcon Minutes: http://lists.w3.org/Archives/Public/www-qa-wg/2004Aug/0140.html (draft)

Minutes: 

 * F2F preparations:
	It is suggested that WG participants share the info on which hotel
they're going to for our upcoming F2F meeting in Reading; also, it would
be worth investigating whether we'll have teleconferences facilities in
the meeting room. An agenda is needed.	

 Lynne: regarding F2F
 ... logistics available and linked from WG home page
 ... still need to work out an agenda
 Dimitris: I should be able to attend the F2F in Reading
 LR: when you know where you'll be staying at, let the WG know
 ... two options I've been exploring: Holiday Inn and Royal Country
 ... Hopefully Karl and Olivier will prepare the agenda
 David: if I know the time slots, I can probably see whether I could attend on the phone
 ... 28 Oct looks unlikely though
 LR: if there is something specific you would be part of, let us know so that we can accomodate the agenda
 ... needs to check with Andrew whether there will be teleconf facilities

* QA WG comments on WebArch
	DHM will send as feedback on WebArch from the WG: that the wording of
the GP on extensibility should be changed to make it clear that breaking
conformance is not an option; suggesting to link to SpecGL and
Variability in Spec; asking for clarification on what the "experience
shows" is about

 LR: comments need to be sent this week
 DHM: (summarizing the thread)
 DM: here is a rough thought on the topic:
 ... a WG may have 3 different intents wrt evolution:
 ... * it should not evolved, e.g. XML 1.0;
 ... * using extensions as explorations of a future version of a spec
 ... * adaptibility mechanism to fit some practical applications around a common core
 ... The TAG wording seems to focus narrowly on the 2nd case, e.g. XPath
 ... so that doesn't match all the different scenarios
 DHM: I'm not sure that's what the TAG really says
 ... looks like they rather say "provide extensibility hooks to allow smoother evolution"
 DM: slicing it in a different way
 ... let's talk about versions
 ... e.g. suppose that XML Core WG wants that there is only one version of XML, XML 1.0
 ... they rely on Unicode to define the allowed character set, so in this sense, XML is extended without any action from the WG
 PC: but is that really an extensibility option? I don't think so
 DM: so you agree with me that the TAG def of extensibility leads to a wrong direction
 ... extensibility is applicable to one specific version of spec
 ... how much it affects interoperability, vs how much adaptibility it allows [?]
 LR: I agree with KD and LH on their remarks of the first GP of WebArch
 ... ie the 2nd SHOULD NOT should be MUST NOT
 DHM: (trying to give why he thinks this has been set as SHOULD NOT vs MUST NOT
 ... conformance vs semantics, maybe?
 ... also, WebArch tries to avoid using MUST
 LR: I understand that, but that should be clarified
 agreement that there is an issue here, esp. about the fact that an extension can interfer with conformance
 LR: also problem with what the "experience shows" is really about
 ... but otherwise, don't see that there is so mucn an issue about looseness 
 DHM: anything else that I should mention in my comments?
 LR: linking to our documents as suggested by Karl

* Issue about " Quality Control" section in Spec GL
	Karl had noted that it wasn't possible to determine as an external
reviewer whether the WG is practicing a quality control on its spec
development. While he suggested this may mean we should define profiles
for SpecGl, there was agreement this would make SpecGL more complex than
it needs; more generally, it was agreed that our GP and principles
should be assessable with visible results, and that this should be a new
issue for our documents.
	As to how to deal with the specific section, it should be either
reworded in that direction, or be downgraded to a GP or Principle.

 LR: issue with section of SpecGL about Quality Control
 DHM: (summarizing KD point)
 LR: the doc is explicitly not specifically targetted to external
reviewser
 ... don't want to make it more complex for that
 DHM: also, maybe we should not require a process (do something)
 ... but rather a result (show that you did something)
 ... we need to see whether this can be accomodated
 LR: for the most parts, our GP and principles result into something
that can be checked
 ... the Quality Control section is indeed the only one that braeks that
rule
 ... this is something we should raise as an issue
 ... if we cannot come up with a way to reword this principle, maybe it
should go away as a principle (e.g. as a GP)
 ... hoping we can reword it
 ... the Quality Control section is still the fuzziest one
 ... wouldn't mind if it was turned into a GP, FWIW
 ... issue: look at all the GP and principles to ensure that by
performing them, you actually end up with a visible result
 ... so to summarize, I don't agree with Karl, but I agree that there is
an issue

* Persons to contact to get feedback on new SpecGL
	Karl made a short list of people to contact to ask for feedback on the
new docs; the WG trimmed it down so as not to dilute our requests for
feedback, and Lynne and Dom will contact the shorter list.

 LR: anybody else (not on Karl's list) we should contact?
 DHM: should we actually ask them? Why not wait LC?
 LR: indeed, they may become bothered at reviewing our doc over and over
 ... but it would probably make sense to address a specific mail to these people
 ... (also, andrew Eisenberg)
 DHM: agree on the principle, but should make clear that we're planning to go to LC in the upcoming weeks
 ... don't think we should contact the TAG now
 LR: Jeremy is probably alraedy awayre
 ... and DanC's probably too busy
 ... I guess I should contact Andrew Eisenberg
 DHM: Susan probably already knows through spec-prod
 ... I'll contact Jeremy informally

*  GEO Tests threads on www-qa
	
 LR: discussion between Jeremy and Al on applying testGL to GEO tests
 ... Al disagreeing with the principles
PC: I'll take a look and see if we can take any issues out of it

* Next meeting
	Next meeting will be as scheduled next week; nobody announced regrets.
-- 
Dominique Hazaël-Massieux - http://www.w3.org/People/Dom/
W3C/ERCIM
mailto:dom@w3.org