QA Working Group Teleconference

F2F in Dublin

dom: anybody prefering July vs August?

patrick: I don't have a preference
... equivalent wrt hosting too

dimitris: I'd prefer Aug 9-11

dom: so would I

tim: this would impact our timetable wrt end of charter, since we're moving it back

RESOLUTION: we're moving the meeting to Aug 9 til Aug 11

SpecGL issues

Issue 995 "old potential issues"

Lofton: I thought I had to review them by April 4th
... I haven't done it yet

Issue 983

karl: need of an example demonstrating the use of an ICS as part of a conformance claim
... didn't find exact examples
... but ATAG 1.0 is close enough
... since they require to explain what implementers have not implemented
... so we can either change our requirements so that ATAG works
... or make it an example with reservations
... UAAG is another similar example
... see http://lists.w3.org/Archives/Public/www-qa-wg/2005Mar/0104.html

dom: we could use them as "could be better" examples

karl: but is what we recommend really better?
... isn't it enough to simply list what hasn't been implemented?

dom: I think the point of the GP is to have a formalized way to express these data
... not a strong supporter of this GP, so probably not the best input

patrick: I think it is useful to require it, it's useful information

karl: so, let's add them as 2 examples, saying we recommend to require a link to the fulfilled ICS
... and that the examples are the 1st steps into that direction

lynne: are we going to include our own example, with SpecGL?
... this would be an exact example
... linking to a completed ICS of SpecGL for SpecGL, and showing it as an example of a claim using an ICS

karl: we need a volunteer to actually fill the ICS, as acurately as possible
... this can only be done when the spec is edited as final

lynne: depends on when this needs to be done
... can we leave this open until when the doc is ready?
... I can't commit to it now, but maybe later
... that's part of the TAG comments

karl: let's record in the issue that this isn't resolved until the ICS is filled

RESOLUTION: we'll publish a fullfiled ICS for SpecGL and use it as an example for the relevant GP

tim: how does the ATAG conformance levels fit in our SpecGL view?

karl: ATAG's conformance model is pretty complex, using atomic details
... the ICS just needs to respect what the spec says
... so an ATAG ICS can be filled with multiple levels

tim: when claiming conformance to a certain level, you declare N/A for items in other levels
... but sometime, people use it as a way to show where they are in the process of getting to this and this level, plus some other items they may have met
... important as a marketing tool

lynne: when claiming conformance to a level, you should only list the relevant checkpoints in your ICS
... it's up to the spec authors to decide how their ICS is organized wrt profiles, levels, ...
... I don't think there are wrong answers on that topic

tim: just mentioning that often, an ICS can be used outside of a conformance claim
... it may still have value out of a conformance claim, e.g. as a marketing opportunity

Lofton: I still dispute that using an ICS is in fact still making a claim
... we shouldn't bother about how marketing uses an ICS

[same ole discussion restarting]

RESOLUTION: We'll also use ATAG and UAAG as part of our example for including an ICS as part of a conformance claim (issue 983)

Issue 1041 - Conformance is not a yes/no proposition (wrt filling an ICS)

Issue 1058 New Numbering Structure

karl: I've integrated the new numbering scheme in the Editors version of SpecGL
... unless anybody objects, we should close that issue

dom: sounds good to me

lynne: to me too

karl: dom, can you use an XSLT to create the ToC for the document?

dom: sure

karl: also, we'll need a new ICS for the document

<scribe> ACTION: dom to update his XSLT stylesheets to create a ToC and ICS for the editors version [recorded in http://www.w3.org/2005/03/21-qa-minutes.html#action01]

karl: also, should we publish the correspondance table as an appendix?
... so that people used to the old numbering scheme can find their ways

dom: I don't think it's useful to have in the document
... what about linking it from the Changelog?

<scribe> ACTION: karl to create the correspondance table between old and new numbering in QA space and link it from changelog [recorded in http://www.w3.org/2005/03/21-qa-minutes.html#action02]

RESOLUTION: issue 1058 is accepted as closed

Issue 1041 - Conformance is not a yes/no proposition (wrt filling an ICS)

-> http://lists.w3.org/Archives/Public/www-qa-wg/2005Mar/0085 Lynne's proposal

karl: we need to get consensus on this issue

lynne: I agreed with some part of Lofton's comments wrt implemented/tested

lofton: I still think it's absurd to work around conformance claims and ICS

[and again, discussion on whether an ICS can be used for anything else but conforming claim]

scribe: I disagree with our approach, and I don't think what lynne suggested reflects what was decided

karl: I'm satisfied with Lynne's proposal, but would rather get real consensus on this

-> http://lists.w3.org/Archives/Public/www-qa-wg/2005Mar/0091.html lofton's counter proposal

lofton: I think Lynne's proposal is still better than nothing

RESOLUTION: issue 1041 is resolved per Lynne's amended proposal

<scribe> ACTION: karl to incorporate last changes from Lynne on ICS definition http://lists.w3.org/Archives/Public/www-qa-wg/2005Mar/0092.html [recorded in http://www.w3.org/2005/03/21-qa-minutes.html#action03]

Issue 1059 Specification definition

-> http://www.w3.org/Bugs/Public/show_bug.cgi?id=1059 Issue 1059

<lofton> http://www.w3.org/Bugs/Public/buglist.cgi?query_format=specific&order=relevance+desc&bug_status=__open__&product=QA&content

karl: reading old and new definition
... reading old and new definition]

patrick, dave, tim: the ISO one reads better

<lynne> specification - document that prescribes technical requirements to be fulfilled by a product, process or service

<lynne> From ISO Guide 2-4

RESOLUTION: to use ISO's definition of "specification" for SpecGL and ViS, referencing ISO Guide 2-4

ACTION karl to update SpecGL -inline and glossary- and QA Glossary with ISO's def of spec, plus add a ref to it

issue 1160 rewording GP 2.3 to be less workflow oriented

-> http://www.w3.org/Bugs/Public/show_bug.cgi?id=1160 rewording GP 2.3 to be less workflow oriented

Dave: there was some discussion on the verbiage used for the good practice

karl: Agree with Dom that the proposed sentence was a bit long

-> http://www.w3.org/mid/OFDFB3C373.424196FF-ON85256FCB.0057CB2C@lotus.com Dave's latest proposal

dave: sent a proposal this morning in 9 (long) words "When imposing requirements by normative references, anticipate conformance dependencies. "

karl: what do you mean by "anticipate"?

dave: that you explains somehwere in your specs how the conformance models interact

karl: I think the 1st part is fine
... but I don't think the 2nd one is testable

dom: we really need this to be about results in the spec vs process to go there

dave: issue is whether "provide" is strong enough
... what about "address"
... ?
... this makes it clear you intend to see something in the verbiage

RESOLUTION: ex GP 2.3B reworded in "When imposing requirements by normative references, address conformance dependencies. "

RESOLUTION: we adopt the rest of Dave's proposal as is for issue 1160

next meeting

dom: chaired by Patrick, on Apr 4th; patrick will need to send the agenda since neither karl nor I are goign to be aroudn before that
... (no meeting next week)

karl: I'll try to do as much work as possible this week on the document
... but we need to work on the ICS

ACTION dimitris to develop a detailed implementation report for SpecGL

<scribe> ACTION: dimitris to develop a detailed implementation report for SpecGL [recorded in http://www.w3.org/2005/03/21-qa-minutes.html#action04]