Minutes for QA WG F2F, Tue Jun 15 afternoon

Scribe: Dominique Hazaël-Massieux

Attendees:
(PC) Patrick Curran (Sun Microsystems)
(KD) Karl Dubost (W3C, WG co-chair)
(DH) Dominique Hazaël-Massieux (W3C)
(LH) Lofton Henderson (CGMO - WG co-chair)
(LR) Lynne Rosenthal (NIST - IG co-chair)
(MS) Mark Skall (NIST)

Regrets:

(DD) Dimitris Dimitriadis (Ontologicon)
(AT) Andrew Thackrah (Open Group)
(VV) Vanitha Venkatraman (Sun Microsystems)

Summary of new action items:
AI-20040615-11 LR to align the terms between SpecGL CR, SpecLite and the
Central glossary by 2004-06-25
AI-20040615-12 KD to turn "too authoritarian and fierce" bullet in a
positive way by 2004-06-16
AI-20040615-13 DHM to create a wiki topic on the qualities of a good
test case (or a topic for each of the concept listed on the paragraph)
AI-20040615-14 DHM to fix QAH with regard to usage of "staff"
AI-20040615-15 LH to look at the "quality *" in QAH in details by
2004-07-16

Agenda:
* QA Glossary
* QAH issues processing

--------------------------
--- QA Glossary Topics ---
Summary: the QA WG confirmed his resolution of having a central glossary
and a glossary per-guidelines document, and set up a plan to make this
happen in practice

LH: only one of the def of SpecGL is in the QA Glossary
... I sent suggestions to add the terms in the QA Glossary
... As a general principle, we ought to make sure that the QA Glossary
is consistent with it
... this has been a cause of concerns in the CR reviews
MS: are we abandonning the idea of making the QA glossary much more
comprehensive?
LH: in Seattle, we resolved that the QA Glossary ought to be
comprehensive
... and any definition in the QA Glossary can also be in the GL
... (with possibly more context-dependent stuff in the GL)
MS: so, the QA WG glossary will be shorter?
LH: it will have more terms, with with more terse definitions
PC: I still have a concern with definitions being in 2 places
MS: but the definitions in the GL can be a proper subset of the general
glossary
PC: but why have it in 2 places then?
LH: it's easier to have the definitions directly in the GL
... but still good to have a comprehensive glossary
LR: we have a sync problem
KD: why do you want to have the glossary in the GL?
LR: so that you can everything you need in the document
... not a link away
KD: a possible solution would be that each time an editor wants to add a
term in the glossary of the document, he copies it from the QA glossary
MS: but having it in several places, you can have it a more appropriate
definition in the local one
LR: but then you risk to confuse people if we're using different
definitions across our documents
MS: just keep the generic definition in the 1st paragraph, and add
context-dependent definition afterwards
LR: if in SpecGL, I can embed the definition in the text, and then link
it to the Glossary, that's fine with me
DHM: I think having the glossary embedded in the spec is good wrt
usability
... it allows better readibility when jumping into the document
KD: agreed, but having a copy of the glossary inside the document would
accomplish this
... but this leaves up the question of "extended" local definition
LR: a dictionary has often several definitions
... the glossary could do that
MS: I rather see a terse definition in the main glossary
... so that you keep consistency
DHM: but you could keep the consistency at the main glossary level

RESOLVED: the QA Glossary has the official definitions ; the GL will use
that defintion (copied), and can add addition context-specific
elaboration of the definition
(basically, a refinement of the resolution of issue 19)
PC: I fear this will leave to divergence
... but I can live with it
... we live in an hypertext world, I think having all in a different
page is OK
LH: this topic was one of our big fight back in Tokyo
... UAAG is an example of that resolution
PC: I still don't know how this will work in practice
... how do we make sure that our GL glossary is indeed in sync with the
central one?
LH: when we go to publish a GL doc, we need to take a look at the local
glossary and the QA glossary
DHM: I agree with PC that I don't know how we'll indeed ensure this
... given we have resolved this for more than 2 years
... and haven't been doing so
LH: but getting read of the central glossary doesn't make the sync
problem go away
... you still need to ensure consistency across documents anyway
DHM: but this doesn't work very well with the time pressure at
publication time
... ideally each time an editor add a definition in the doc's glossary
... he should ask to add the term to the main glossary
LH: agreed, but we still need to make sure at publication request time
that we're indeed in sync

RESOLVED: Lynne will ensure at publicaton time that there is consistency
between the local and the central glossary
ACTION:
AI-20040615-11 LR to align the terms between SpecGL CR, SpecLite and the
Central glossary by 2004-06-25

---------------------------------
--- QAH Issues and next steps ---
Summary: the WG resolved most of the open issues on QAH; the following
issues are still pending though: licensing terms for test materials
(waiting for input from the W3C Advisory Board), fuzzy usage of the word
"quality" in QAH

QAH Issues list: http://www.w3.org/QA/WG/2004/06/QAH-issues.html

*** "Don't use 'IPR', but 'legal issues'" *** 
LH: the points made in the referenced documents made sense to me
... I'm OK with changing this
PC: I don't see the point ; IPR is widely used in the industry
... legal issues looks as fuzzy if not more than IPR
... since this is not a lawyer doc anyway, I don't see a compelling
reason to change
LR: plus, W3C uses "IPR" too, right?
KD: [checking]
... "Intellectual rights" seems to be the proper term used inside W3C
... 'IPR' is used in many places inside W3C
DHM: but in official places?
KD: it's used in the process document, at least (>10 times)

RESOLVED: reviewing our usage of IPR in QAH, we clarified IPR by more
specific terms (details of implementation below)

Going through "IPR" occurences in current QAH:
* "IPR and License issues"-> "licensing issues"
* "Test materials license and other IPR issues"
-> "test material licensing isuses"
* "IPR issues" -> "licensing issues"
* "the IPR and staff issues" -> "the licensing and staff issues"
* "Sort out IPR issues with the external party " -> "licensing issues"
* "W3C IPR license" -> "W3C license"
* decisions about IPR and test materials. -> "licensing of test
materials"

*** issue JC-2 ***
"QAH takes this approach for several reasons
... too authoritarian and fierce"
RESOLVED: we'll try to turn it in a positive way ; if we don't find
something, we'll just remove it
AI-20040615-12 KD to turn "too authoritarian and fierce" bullet in a
positive way by 2004-06-16

*** editorial issues serie ***
JC-3, 5 to 11, 16  (mostly editorial)
RESOLVED: agreed with proposed resolution

*** issue JC-13 ***
"copy&paste issues with document licenses"
RESOLVED: legal issue wrt documentation license is pending, waiting for
input from the Advisory Board
LH: will this be available before next publication late summer?
DHM: absolutely
... then, the question of whether we agree with the AB is a different
topic

*** issue JC-14 ***
RESOLVED: agreed in principle; since TestGL is not ready yet, we'll
create a Wiki topic on the relevant points and link it from the handbook
until testgl is ready
ACTION:
AI-20040615-13 DHM to create a wiki topic on the qualities of a good
test case (or a topic for each of the concept listed on the paragraph)

*** issue JC-15 ***
"TM expanded, but been used before"
RESOLVED: agreed; and as a general rule, we'll try to eliminate as much
use of abbreviations as possible and reasonnable

*** issue JC-18 ***
"QA IG's  deliverables"
RESOLVED: agreed -> "in order to effectively fulfill QA IG reviews"

*** issue JC-19 ***
"QA Framework primer unnecessary"
PC: it looks a bit redundant as it stands right now
... it would be less so if it were inline, or as a checklist
LH: but some of these doesn't fit above
... it was in the intro doc, and the content of the handbook is more
opsgl
... this is an umbrella to handbook, specgl and testgl
LR: nobody seems to have a strong opinion at this stage
... somebody should look at how to handle the primer, what it should be,
etc.
... I think it should be a separate 3 page overview of all our documents
... I don't think it makes sense to have it as an appendix to the
document

RESOLVED: we'll move the primer from the qa handbook to /QA/ web page
(QA Framework roadmap)

*** issue DHM-1 ***
sub issue 1): dropped
sub issue 2)-> AI-20040615-14 DHM to fix QAH with regard to usage of
"staff"
sub issue 3): agreed with proposed resolution

*** issue DHM-2 ***
RESOLVED: pending, waiting for AB policy proposal

*** issue JC-4, 17 ***
LH: I agree that there is some preponderous emphasis on testability
PC: there are indeed some QA aspects that we don't address
... performance, document quality, usability
LH: but should we rename QA Framework to test and testability framework?
PC: I don't think so, we just need to include some rewording as proposed
LH: what about s/QA Projects/Test materials/ ?
DHM: in editor version, "test task force moderator"
LH: we've been using "QA moderator" too
DHM: but in this case, it seems to be limited to testing
... and in current WGs, there are actual test suite development leads,
but no QA moderators
KD: "moderator" is not a good term, it has negative aspects; what about
"QA lead"
LH: we could mention that there are several potential roles behind this
QA lead
... which could be distributed among several people
... or we could just drop the label
LR: you're right
... in the examples, we can take the terms actually used in given
Working Groups
LH: we've been using "QA Point of contact" in the QA PD template
DHM: what about the original text for this issue (QA projects)
LH: we could change "Test TF moderator" to QA PoC
LR: as of today, there are actual Test TF leads, we shouldn't miss them
behind the QA PoC
DHM: what about renaming "Test TF moderator" to QA PoC? The Test TF lead
doesn't have much more to read than any other Test TF participant
LH: and let's change "QA Projects" to "test and other QA-related
activities"
LR: agreed
RESOLVED: agreed to add words to scope saying that there is an emphasis
on testability ; not renaming QA Framework ; changing "QA Projects" to
"test and other QA-related activities"

*** "QA deliverables" issue ***
fuzzy terminology "QA deliverables" et al (e.g. quality practices)
LH: there was no clear agreement on MS's proposals
DHM: I note that our current QAPD is only about test material
development
... it has nothing about other QA-aspects
(general discussion about scope of our documents and our work)
PC: if our primary focus for now is about spec testability, conformance
testing, we should state it
(more discussion about scope of QA, QA Framework, meaning of quality,
...)
ACTION:
AI-20040615-15 LH to look at the "quality *" in details by 2004-07-16

Dom
-- 
Dominique Hazaël-Massieux - http://www.w3.org/People/Dom/
W3C/ERCIM
mailto:dom@w3.org