DRAFT minutes 2002-07-03 telcon minutes

All,

please sanity check.

---

QA Working Group Teleconference
Friday, 3-July-2002
--
Scribe: [Dimitris Dimitriadis]

Attendees:
(dd) Dimitris Dimitriadis (Ontologicon)
(KD) Karl Dubost (W3C, WG co-chair)
(KG) Kirill Gavrylyuk (Microsoft)
(DH) Dominique Hazael-Massieux (W3C - Webmaster)
(LH) Lofton Henderson (CGMO - WG co-chair)
(MS) Mark Skall (NIST)

Regrets:
(JM) Jack Morrison (Sun)
(LR) Lynne Rosenthal (NIST - IG co-chair)
(SM) Sandra Martinez (NIST)

Absent:
(PF) Peter Fawcett (Real Networks)
(AT) Andrew Thackrah (Open Group)
(OT) Olivier Thereaux (W3C - systems)

Summary of New Action Items:
2002-07-03-1: David Marston to send proposed wording for splitting 
checkpoint 4.6 (DM)
2002-07-03-2: Define borderline between in and out of scope (spec 
interdependencies) (KG)
2002-07-03-3: Check Test GL editor's draft for checkpoint verifiability 
(MS)
2002-07-03-4: Editors' action item to look at splitting Guideline 4, 
already conatins 13 checkpoints (16 with additions from DM)
2002-07-03-5: Editors' action item to expand checkpoints pertaining to 
tests (test validity, test case and test suite management issues; will 
appear in next editors' draft of Test GL)

Previous Telcon Minutes:
http://lists.w3.org/Archives/Public/www-qa-wg/2002Jul/0000.html

Minutes:

Roll Call (results above)
Also present: (DM) David Marston, guest

Tokyo preliminary attendace call
dd - to be decided
KD - yes
KG - so far yes
DH - yes
LH - yes
LR - probably
MS - probably
SM - probably

Spec Extech Case Studies

resolution: Karl has no special requirements, KG and KD are going to 
collaborate on Kirill's generation of SOAP and Karl's submission of it 
to the XML WG

Continued Test GL chackpoint review
3.3 finished last telcon

Test Framework:

(KG) Quick review of some of the issues touched on last telcon to see if 
they can be closed

1.1 Create list of all specifications used for reference. Agreed to be 
those being tested and those being referenced. have the WG comments been 
adressed? (no objections) closed for now, if anyone sees anything, let's 
raise the issue again

1.2 Extract test assertions. (interdependencies between specs, XSLT can 
include tests from say XPath, whereas XML cannot).

(DM) We need to have a test case be able to cite a test assertions. Not 
sure who is doing the extracting.

(KG) The TS development

(DM) probably applied indirectly, then

(KG) If you rely on the spec and only on the spec format, do you need to 
include tests on the particular format? Are they out of scope?

(DM) In XSLT XPath, you will find examples that are in scope

(KG) Resolution to define the borderline between in scope and out of 
scope [Action item]

1.3 has not been adressed yet

Guideline 4

(Dimitris' comments in last editor's draft discussed) Make clearer the 
distinction between choosing existing framework, if applicable, or using 
checkpoint to produce suitable framework.

Checkpoint 4.1 will be changed according to comment (no objections)

4.2 (MS) How do you verify that this checkpoint has been met? One way is 
to execute the framework on a number of platforms in order to have it 
verifiable.

(KG) Not sure how to solve it, will record it as an issue. Applicable 
only for non-existing frameworks, if they exist it will have been checked

4.3 Similar discussion: does platform independent imply it runs on all 
platforms? Not necessarily, something may be written for a piece of 
machinery or software which is not conformant (eg. old browsers)

(MS) Generic problem we will have on verifiable checkpoints. Volounteers 
to go through the document and check all checkpoints for verifiability. 
Will look at Lofton's 20020702 email.

(DH) Do we need to state something about complete platform independence, 
or only as independent as the application of the specification itself?

4.4 Unverifiable. Wording should be changed to identify the adding 
mechanism (dd & MS)

4.5 Similar to the above

4.6 (dd) Add "for tests" in wording

(DM) slpit into four:
	with regard to the framework and collecting a set of tests, 4.6 
splits into spec errata and spec version
	with regard to versioning of the test suite, that's a different 
checkpoint which isn't there
	also, each test can be marked wrt spec version and errata level

4.7 (KG) Action item: make the guideline wording in synch with the Spec 
Guideline wording

4.8 (DM) maybe this is where the question of variability comes in

(LH) You can either consolidate 4.7, 4.8 and 4.9 into a single 
checkpoint that talks about accomodating variability, or you can have a 
new guideline to accomodate variability

(KG) If the priority is the same for all of them, there is no reason to 
itemize them

(LH) Guideline 4 needs to be altered in any case, we already have 13 
checkpoint, maybe we should take an action item to split it into more 
guidelines

categories: variability, test automation (def as part of the framework 
where you apply a collection of tests), case management (choosing 
relevant tests for particular implementations)

4.9 OK as is

4.10 (dd) change wording to platform and product independent, 
verifiability issues applies here too

(KG) recorded as an issue

4.11 (LH) Who does the framework allow documentation?

4.12 (KG) More specific definition of test case management required

(LH) Maybe this together with 4.6 form a separate guideline

(KG) Moved to editors' group discussion

4.13 Further investigated and brought into synch with wording of Spec 
Guidelines

(dd) We need to discuss test validity somewhere

Next Telcon

(LH) Start the next Telcon with some Spec GL issues in order to be able 
to go public. Next teclon is the next regular scheduled one, July 10. We 
do not anticipate to have one July 17

Special telcon on Test GL in August after editor's draft July 19

Adjourn
Meeting adjourned at 11.30 EDT

Received on Wednesday, 3 July 2002 12:28:04 UTC