Minutes of QA Working Group Teleconference, 2002-12-02 [FINAL]

QA Working Group Teleconference
Monday, 02-December-2002
--
Scribe: [Dimitris Dimitriadis]

Attendees:
(dd) Dimitris Dimitriadis (Ontologicon)
(KD) Karl Dubost (W3C, WG co-chair)
(PF) Peter Fawcett (RealNetworks)
(DH) Dominique Hazaël-Massieux (W3C)
(LH) Lofton Henderson (CGMO - WG co-chair)
(SM) Sandra Martinez (NIST)
(LR) Lynne Rosenthal (NIST - IG co-chair)
(MS) Mark Skall (NIST)
(DM) David Marston, (IBM)

Regrets:
(JR) JohnRobert Gardner (Sun)

Absent:
(KG) Kirill Gavrylyuk (Microsoft)
(AT) Andrew Thackrah (Open Group)

Agenda: 
[http://lists.w3.org/Archives/Public/www-qa-wg/2002Nov/0111.html] 
(draft)
Previous Telcon Minutes: 
[http://lists.w3.org/Archives/Public/www-qa-wg/2002Dec/0003.html]

Minutes:

(lh) second item from agenda is dropped

(lh) spec gl issues and topics

(lr) I'll go through in the same order as I sent out

(lh) replies on reference 4 from me and mark

(lh) in december archives

(lr) two levels to think about

(lr) what types of changes to we need to make ono CP in order to ensure 
that it is clear and understood what we want and is expected in order 
to satisfgy it

(lr) also, have we gone too far or not far enough?

(lr) other point, to think about the CK applying not only to Technical 
recommendations but also how it applies to Guidelie specs, such as QA 
and WAI

(lr) CK is OK for XML and Schema (tech spec) not sure when applying to 
guidelines

(lr) second issue: ensure that the Spec GL conforms to itself

(lr) was exercise for myself and Dom, done by Mark (good reading)

(lr) when a CK does not apply should the Spec indicate that this and 
that does not apply? Should it be required? Is the appropriate place in 
an ICS (Implementation Conformance Statement)?

(ms) many times I had to reread SpecGL

(ms) if we want to determine conformance, it should be easy to have the 
implementation report on when a particular CK does not apply

(ms) maybe an appendix, if not an ICS

(lh) ICS included is an ICS for documenting conformance, but not to 
conformance to Spec GL

(lh) I sympathize with what you say abaout discoverability, the problem 
I have is, if we require too much overhead on Spec authors in 
achieveing conformance to Spec GL (documenting in their specs do not 
apply), I'm afraid we're going to generate lots of resistence to using 
Spec GL

(lh) I'd like to figure out a way to do this lighter

(lh) When they fill out the ICS, they can document there what is 
applicable and what not, hesitant to force too much structure though

(ms) some CK specifically say when they do not apply, others don't. we 
should be consistent

(ms) was not aware of the discussion of burdensome procedure

(lh) talk on Chairs list email archive, about XForms 1.0 CR request 
contained talk on burdensome procedure, not done in general

(lh) I18N wrote their mandatoriness into their Spec

(lh) some informative text to the effect that we've anticipated 
comments on applicability, if it is not possible to documents reasons 
for inapplicability, please
document

(lh) all the non-applicables can not be documented

(lr) bottom line: our Spec GL will have some ICS, we will require as 
part of our conformance statement that the ICS has to be completed and 
that we'lll make sure that the ICS makes it clear that all CK say 
something about applicability

(lr) 2.1: identify all classes of products

(lr) within Spec GL it isn't clear what our class of product is, we 
need agreement on that

(lr) we believe that W3C understand what class of product is

(lr) is target a similar as class of products? ('target' read in UUAG)

(lr) one class or two classes?

(dh) W3C Technical reports

(dh) it applies to any

(lh) CUAP: Common User Agent Problems, is a Note W3C has written

(kd) common problems and non implemented features

(lh) is one an example of where Spec GL is not applicable

(lh) fine with Dom's suggestion to leave it simple now and take care of 
any issues later

(lr) our class of product is W3C Technical reports

(dh) do we need to precise that it needs to be on Rec Track or not?

(lr) for now let's leave it simple

(ms) are we going to have a text in Spec GL?

(lr) yes

(lr) GL 2, there's a list of classes of products, should technical 
reports, or specificaqtions, be added to the list?

(lh) Mark believes we should, I think it's not necessary

(lr) so we basically do bullet 1, through the third CK

(lr) the class of pruduct that Spec GL is. is W3C Tech Report. Insofar 
as adding something, it will be a category called Guidelines, as far as 
the list of classes of products, we leave it alone

(lr) 13.2 distinguish normative and informative text to fulfill the CK 
you must dist. nor/inf, does this mean that you have to label each 
section? do you make a general statement up front? are we asking people 
to label everything? is it a burden for people writing specs?

(ms) either label every section, or do a default (everything is 
normative unless otherwise stated)

(lr) we need to clearly understand what is normative and what 
informative

(dh) in the spec, what I have done is that what is normative appears 
with a grey background

(dh) include keywords in the text

(dh) we should not try to enforce much more than that

(dd) not too big a burden to add to have a definite reading of the spec 
in the end

(lr) is ExTech a normative or informative reference? second issue; is 
ExTech itself Normative/informative

(dh) ExTech cannot be normative, it is only informative text, technical 
examples

(dh) there are no requirements in there

(dd) isn't one way to make the reference itself normative?

(ms) 10.3 says make reference to specs on which the spec depends, 
brings on issue to what we mean by depends

(dh) dependence implies normative reference

(lh) split "Reference" section into normative/informative section

(lr) do we in 10.3 want to require that normative reference are in some 
list or table?

(lr) do we need to come up with a better word than depends?

(lr) second point, do we nede to require that they be given in a list?

(dh) we already have a dependece definition

(dh) we need to move it into the definition section

(lh) sandra made a definition of formal descritpions

(lh) are we going to require that there is a normative reference 
list/table?

(lh) add it to the fast discovery lsit

(ms) quikly findable is clearly not measurable, requirements should me 
measurable

(lr) CK 13.4

(ms) why not put it in one place?

(dh) everyone agrees on the list, the goal is more important, the list 
is one way

(lh) should we schedule another telcon for SpecGL?

(lh) would week of the 16th dec give us enough time for seattle F2F?

(lh) will try to use old slot Wed 18th Dec

meeting adjourned 8 minutes past the hour

Received on Monday, 9 December 2002 12:16:25 UTC