minutes of 20040816 telecon

QA Working Group Teleconference
Monday, 16-Aug-2004
Scribe: Lofton

(PC) Patrick Curran (Sun Microsystems)
(DD) Dimitris Dimitriadis (Ontologicon)
(KD) Karl Dubost (W3C, WG co-chair)
(DH) Dominique Hazaƫl-Massieux (W3C)
(LH) Lofton Henderson (CGMO - WG co-chair)
(LR) Lynne Rosenthal (NIST - IG co-chair)

(DM) David Marston (Guest from IBM)

(MS) Mark Skall (NIST)

(AT) Andrew Thackrah (Open Group)

Summary of New Action Items:
AI-20040816-1 -- Dom -- Publish TestGL placeholder [0] -- 20040823
AI-20040816-2 -- Dom -- Email QAWG about stronger connection between norm-lang
definition and conformance designations-- 2004????
AI-20040816-3 -- Karl -- Start email discussion of various aspects of 
norm-lang issue -- 20040816
AI-20040816-4 -- Lynne -- Draft conformance criteria for SpecGL (Principles 
are normative) -- 20040820


Previous Telcon Minutes:


(KD) Next week's telecon (23-aug) will happened as scheduled.

(DH) Takes action item to publish TestGL by Monday (8/23).

     * Topic.  3.) SpecGL Draft. [1]

Summary.  We will target publishing QAH and SpecGL Monday 30 Aug instead of 
Fri 27 Aug.  KD gives progress report on what remains for him to do before 
publication.  A few figures, some bits on PRs and GPs.  All contributions 
requested by end of this week (20 Aug).  Okay?  All -- yes, okay.

(KD) About SpecGL publication... should References section formally include 
all references, those that are given in examples?
(DH) Yes, good idea.

     * Topic.  4.) Normative-language wording -- Terminology section or 
Conformance Clause [2]

Summary.  Issue seems to have three related parts:  1.) we don't conform to 
our own normative-language GP in section A;  2.) the norm-lang GP implies 
that wording like RFC 2119 boilerplate must be in Conf. Clause, instead of 
in a separate section elsewhere;  3.) the normative-language GP in section 
A duplicates stuff in section C, and KD proposes removal from section 
A.  There seemed to be some agreement (maybe not unanimous?) that a link in 
Conf. Clause to (e.g.) RFC2119 boilerplate section elsewhere was acceptable 
solution to #2.  Disagreement about other parts, to be taken to mailing list.

(KD) Normative wording issue [2].  KD proposes to remove GP "how norm lang 
is expressed, in Conf Clause".  And add "link okay" to surviving GP in 
section C.

(DH) Should norm-lang GP be tied more strongly to conf-designations?  Dom 
to write to mailing list about the issue.

(DH) Link issue:  Dom doesn't see the need for a link from Conf. Clause to 
something like a RFC2119 boilerplate clause, which is common and widely 
used in W3C.

(LH) On the other hand, we intend that "everything you need to know about 
conformance to [spec] can be found starting in the Conformance 
Clause".  Which implies that a link (at least) should be required there.

Lots of discussion and some disagreement about various aspects. LH objected 
to removal of norm-lang GP from section A.  Resolved to take it to email 
for further discussion.

(KD) takes AI to write something to mailing list today.

     * Topic.  5.) Modules-Profiles-Levels/DOV [3]

Summary.  General agreement to take Profiles, Modules, Levels (PML) 
explanatory material outside of SpecGL into a referenced /TR/ Note or /QA/ 
article.  But it is pretty short for /TR/ note.  So proposal to combine it 
with DoV stuff.  Issue was then:  /TR/ or /QA/?  And how much coordination 
with future SpecGL publications?  General agreement to combine PML and DoV, 
into a WG Note to be published in /TR/.  It should be coordinated initially 
and published with next SpecGL.  Thereafter it should be re-published with 
SpecGL as needed -- if changes in SpecGL imply changes in Note, or if the 
material (e.g., DoV) is to be more fully developed, then republish 
concurrent with a SpecGL publication.  Otherwise, no need to republish.

(KD)  Issue to take profile, module, level to article in QA 
space.  DH:  Too short to be in TR space by itself.  Or combine with 
something like  DoV and put into /TR/.

(DM)  Feels that it will not still be too short, if we extend the current 
content with DoV interrelationships material?

(LR)  Either way is okay, TR or QA article.

General discussion:  most favor keeping in /TR/ and coordinated with 
SpecGL.  It provides expands on basic definitions, with explanations to 
help understanding SpecGL.  It also provides advanced material for those 
who want to go beyond the simple level that SpecLite will apparently limit 
itself to.  (DM) in a sense it is almost semi-normative.  PML & DoV should 
be combined.

all:  Discussion of whether the PML & DoV stuff needs to be synchronized or 
not.  Agreed that next WD publication must be.  Thereafter, if no changes 
needed, need not republish.  Will happen as WG Note.

(KD)  Summarizing:  1.) The first publication of DoV/PML will be published 
at the same time as next SpecGL  (8/30);  2. The synchronization if needed 
will be discussed after this first publication.

     * Topic.  Should SpecGL have any normative content? [@@]

Summary.  KD proposed [@@] that perhaps SpecGL should not have normative 
content, but should be entirely informative (like QAH).  The group was 
unanimous that it should have conformance model based on normative Principles.

(KD)  Strawpoll on LR's three options [@@]:
1.) no normative content, no conformance definition
2.) conformance definition based on normative Principles
3.) conformance (or not) based solely on presence (or absence) of a 
Conformance Clause.

DH: #2
LH: #2
LR: #2 is best (#3 at least).
PC: #2
DD: #2

(LR) will draft something (by Friday) to implement #2.

     * 7.) Topic. Brief Overview of XQuery Testing TF F2F

Summary.  Xquery/Xpath testing group met at NIST.  Lynne & Mark attended 
and Lynne summarized as follows.

(LR)  Xquery/Xpath testing grp meeting.  LR & MS attended.  LR did quick 
overview of SpecGL etc, TestGL status, etc.  They were real interested in 
both Ops stuff and Spec stuff.  There are interesting questions and issues 
with Xquery stuff -- Xquery is a bunch of modules or pieces, and need to 
explain conformance of pieces in the context of the whole mess.  Very 
receptive.  May have some good feedback from them to us in a couple of 
months.  They were supportive of what QA's doing.

(DM)  [...had some comments to add.]  Spoke to them day before.  Topics 
like DoV.  Discussed with Jim Melton spec editor, on topics like avoiding 
unintended discretion.  Interested by trial implementations, trial test 
suites, etc.  They have rolled back from CR, and are now in WDs again.

(LR) Carmelo said they'd like to meet again periodically at NIST.

Adjourn 12:00 noon EDT.


[0] http://lists.w3.org/Archives/Public/www-qa-wg/2004Aug/0027.html
[1] http://www.w3.org/QA/Group/2004/07/WD-qaframe-spec/
[2] http://lists.w3.org/Archives/Public/www-qa-wg/2004Aug/0015.html
[3] http://lists.w3.org/Archives/Public/www-qa-wg/2004Jul/0019.html
[4] http://lists.w3.org/Archives/Public/www-qa-wg/2004Aug/0031.html

Received on Monday, 23 August 2004 14:40:44 UTC