W3C home > Mailing lists > Public > www-qa@w3.org > April 2002

2002-04-18 QA Working Group Teleconferences Minutes

From: Dominique HazaŽl-Massieux <dom@w3.org>
Date: 26 Apr 2002 11:35:30 -0400
To: www-qa@w3.org
Message-Id: <1019835330.505.22.camel@sirustier>
QA Working Group Teleconference
Thursday, 18-April-2002
Scribe: Dominique Hazael-Massieux

(dd) Dimitris Dimitriadis (Ontologicon)
(PF) Peter Fawcett (RealNetworks)
(KG) Kirill Gavrylyuk (Microsoft)
(DH) Dominique Hazael-Massieux (W3C - Webmaster)
(LH) Lofton Henderson (CGMO - WG co-chair)
(LR) Lynne Rosenthal (NIST - IG co-chair)
(MS) Mark Skall (NIST)
(OT) Olivier Thereaux (W3C - systems)
(SM) Sandra Martinez
(DM) David Marston

(KD) Karl Dubost (W3C, WG co-chair)
(JM) Jack Morrison (Sun)

(DD) Daniel Dardailler (W3C - IG co-chair)
(AT) Andrew Thackrah (Open Group)

Summary of New Action Items: 
A-2002-04-18-1: LR to inform the IG that IG members can participate and
inform the chair if they want; send the preliminary agenda along
A-2002-04-18-2: KD to coordinate the WG response to I18N regarding their
last call of charmod (to be published)

Previous Telcon Minutes:
http://lists.w3.org/Archives/Public/www-qa-wg/2002Apr/0071.html [DRAFT]

Agenda: http://lists.w3.org/Archives/Public/www-qa-wg/2002Apr/0089.html
1.) Roll call
2.) Any WG Logistical Topics
3.) Briefing about Ops-Extech [1] plan [6] (Lofton)
4.) Overview of "spec anatomy" plans [2] (Dimitris)
5.) Continued checkpoint review of Spec Guidelines [3]
         (see also checklist [5], last telcon minutes [4])
6.) Next telcon (regular or extra)
7.) Adjourn
1.) Roll Call
See above

2.) Any WG Logistical Topics
LH: KD is working on the logistics of the f2F
... he sent a preliminary agenda:
.. have a look at that everyone
... if you want to refine/change any topic, send a mail in reply to
ACTION: LR to inform the IG that IG members can participate and inform
the chair if they want
... send the preliminary agenda along
LH: any other topic?
OT: "a week in QA"
http://lists.w3.org/Archives/Public/www-qa/2002Mar/0014.html will start
... other stuff (tm) on the web
... but not real things right now
ACTION: KD to coordinate the WG response to I18N regarding their last
call of char-mod (to be published)

3.) Briefing about Ops-Extech [1] plan [6] (Lofton)
[1] http://www.w3.org/QA/WG/2002/framework-20020405/qaframe-ops-extech
[6] http://lists.w3.org/Archives/Public/www-qa-wg/2002Apr/0088.html

LH: we are planning to publish that in mid-may
LH: wants to brief you about the structure
... and gets endorsement
... but if people feel it's a too short notice, we can move that offline
... it's a companion doc to operational Guidelines (GL)
... we are picking 4 examples of QA experiences: DOM, SVG, XML and
hopefully XSLT
... descriptive overview; you can see that in DOM in the current spec
... [mainly details of
http://lists.w3.org/Archives/Public/www-qa-wg/2002Apr/0088.html ]
... 2 questions:
... - do people endorse the structure?
... - do we endorse publishing in mid-may? knowing that there wouldn't
much progress on the techniques-side
MS: the doc is different from what I had imagined
... this is a valuable exercise for us
... but why should we publish this for the community?
... Wouldn't it better to show how this should have been done?
LH: one part is that we don't have examples of how it should have been
... the guidelines are just about new groups, or re-chartered groups
... it doesn't mean that the current 35 we have are not QA-aware
... Having hyperlinks allows to inform how other groups did
MS: we might not want to encourage the way these groups have done
lofton, could you ask people to give their name before speaking?
LH: the idea is to match the examples with the guidelines
LH: this is only a FWPD to provide examples to the dry prose of the GL
... My second comment is that I envision that doc evolving
... our idea has been evolving while we were putting it together
... This might become a very different doc
MS: I have no pb with publishing
... but we need to make clear that it's not because it has been done
that way that's it is the right way
LH: I fully agree with that
... at the f2f, we might have some time to look at where it will evolve
... it would be easier
dd: one point to keep in mind: we don't want to include a kind of
exclusive list(?)
LH: please comment further on the list
... otherwise, I'll take this as endorsement to continue with this

4.) Overview of "spec anatomy" plans [2] (Dimitris)
[2] http://lists.w3.org/Archives/Public/www-qa-wg/2002Apr/0075.html
dd about http://lists.w3.org/Archives/Public/www-qa-wg/2002Apr/0075.html
dd: rationale: come to the point where we don't have spec difficult to
... one possibility would be to create a XML Schema to help 
... the HTML version would be produced from the XML version
... that would only be one of the various output of the XML doc
... we would have a much better control on the content
... in DOM, we create the test suite directly from the spec through a
markup language
... the test case become normative in a sense (because it's generated
from the spec)
??: would people accept the XML version as being equally normative?
dd: we could make the XML+XSLT the normative version, but that might be
troublesome for the users
LH: David, XSLT has some history with this issue, isn't it?
DM: not really at this level of abstraction,
... at the workshop a year ago, I wanted the XML version to be the
normative version, but there was some gasp from the team members
... but we weren't at that level
LH: let's not debate this issue in too much details
... let's take in on list
... what do we intend to have in the spec GL doc?
dd: the 1st thing I want to put in this doc is ??
... and then how such a doc with a schema could be use for test
DH: do you know about spec-prod
dd: no, I don't
... will look at that
... benefits of migrating in this direction: better testability, ...
LR: what do you mean by smokelist?
dd: you enumerate each feature and you can deduce test assertions from
the list
LH: in SVG, we went trhough and identify each feature and attached a
test to it
dd: we also in // have to look at what it would be the cost to the WG

5.) Continued checkpoint review of Spec Guidelines [3]
         (see also checklist [5], last telcon minutes [4])
[3] http://www.w3.org/QA/WG/2002/framework-20020405/qaframe-spec
[4] http://lists.w3.org/Archives/Public/www-qa-wg/2002Apr/0071.html
[5] http://www.w3.org/QA/WG/2002/framework-20020405/Spec-checklist

LR: I'll try to incorporate all the comments for next week in the
editors' drafts
... we stopped at GL3
... Checkpoint (CP) 1: when you write a doc, what kind of thing are you
speaking about?
... for instance, the conformance is not the same for an authoring tool
and a [?]
... "classes of product" is the term we're using in OASIS
MS: there are 2 things we can do for this GL: either we define an
exhaustive list,
... or we give just examples in an open-ended list
... I'd prefer the second solution
LH: DM mentionned 4 levels of abstraction
DM: the highest level is the spec itself: API, content, or protocol
... and then every spec affects several classes of product
MS: the risk of introducing a different term through a taxonomy is to
confuse the community
... we don't want to confuse people
DM: my answer would be to qualify by subclasses: a generator is an
implementation of a content producer, and the spec just defines what any
producer has to do.

... we could say that we're trying to make an exhaustive list
LH: MS, could you send us a little write-up?
MS: I volunteer
LR: I think we agree you need to identify what classes of product
... and about CP2 and 3
... any suggestion on any other CP?
... re: GL4 
... various ways of splitting a spec
... I use the term modularization and profiling
... to encourage their use and to define them
MS: one comment: 2 ways of dividing a spec -> profile and level
... profile: geared to different consumers
... level: geared to different levels of implementation
LR: levels seemed confusing
PF: in SMIL2, you have modules
... animation and timing modules used to animate SVG for instance
... but there are levels in each module (essential vs cool features)
MS?: I think adding the concept of the level is essential
DM: in addition to PF examples, for CP 4.1, we cannot choose between a/
and b/
... this adds complexity
... maybe we should say "try not to have a too complex mix of
LH: SVG 1.1 defines a modularization; and there are profiles subset of
each modules (tiny and basic)
??: what do we say? profile+module make sense, but not levels+modules?
PF: actually, SMIL2.0 uses the 3: modules (timing, ...), levels
(regular, extended animation), and profiles defined at the end of the
DM: aren't profile the way you said it) just a macro for this
module at this level + that module at that level?
LR: "levels" is confusing
... levels of conformance and levels for dividing a tech are related
... but they are not the same
... any idea about how to do this well?
... we are discussing GL5 at the same time BTW
... do we agree on adding "levels" in ways of dividing?
... no disagreement
LR: any comment on GL5?
DM: CP5.3, it might be allowable for a spec to recommend a verbiage
LR: isn't that CP5.2?
LH: about CP5.1, in some cases it might too productive, in some cases 
too limited
... it might be a good exercise to look at chapter 3 in UAAG
... there is a concept in there about conditionnality
LR: that's flavour (GL2)

6.) Next telcon (regular or extra)

LH: if we plan to publish mid-may, do we need more extra telecon about
... we could make it optional
LR: I don't really need one
LH: for GL 1 to 10 are yours, are you confortable with them?
LR: yes
LH: what about dd?
... re GL11?
dd: monday with editor's draft, we'll be able to decide
LH: ok, we'll wait monday to decide if we need an extra teleconf

Decision will be taken by Tuesday EOB

7.) Adjourn
Dominique HazaŽl-Massieux - http://www.w3.org/People/Dom/
W3C's Webmaster
Received on Friday, 26 April 2002 11:38:41 UTC

This archive was generated by hypermail 2.4.0 : Friday, 17 January 2020 22:43:19 UTC