Final minutes of 16 December 2002 QAWG Teleconference

Final Minutes:

QA Working Group Teleconference
Monday, 16-December-2002
Scribe: Andrew Thackrah


(DD) Dimitris Dimitriadis (Ontologicon)
(PF) Peter Fawcett (RealNetworks)
(KG) Kirill Gavrylyuk (Microsoft)
(LH) Lofton Henderson (CGMO - WG co-chair)
(DM) David Marston (Guest)
(LR) Lynne Rosenthal (NIST - IG co-chair)
(MS) Mark Skall (NIST)
(AT) Andrew Thackrah (The Open Group)


(KD) Karl Dubost (W3C, WG co-chair)
(DH) Dominique HazaŽl-Massieux (W3C)
(JR) John Robert Gardner (Sun)
(SM) Sandra Martinez (NIST)

Summary of new Action Items:

AI-2002-1216-1 LR to contact Olivier, Janet & Judy to attempt
to schedule an expert/brainstorm meeting for the Seattle f2f meeting

AI-2002-1216-2 LH to propose a rephrasing of OpsGL CKPT 5.3 (prior to Seattle f2f)

AI-2002-1216-3 MS to provide assertion examples from SpecGL for one guideline
of his choice. (prior to Seattle f2f)

AI-2002-1216-4 KG to provide links to requirements/assertions in XForms, SOAP
and maybe others for ExTech.


Previous Telcon Minutes:


1. Roll Call (1100 EST), membership

See above

2. Logistical topics

- Seattle f2f, telcon logistics:

[ QAWG is asked to inform KD of any requirements they may have regarding the provision
  of telcon facilities at the meeting]

 LH - How much advance notice do you (KD) need for facilities

 KG - I'll set things up first week Jan. Just in case, I'll set up Zakim but there
       should be a Microsoft number that WG can dial to.

 DD - There may be a problem for me dialling long distance

 LH - We may need a Zakim dial-out backup

 KG - The phone in the meeting room has telcon function.

- Seattle f2f, agenda details:

 [ QAWG discusses the timetable. LH suggests 8.30 to 5 ]

 KG - 8.30 - 5 each day works for me
  [No objections]

 LH - [goes over proposed agenda ] We already have 3.25 days scheduled
  We could throw out .5 day reserve (item 6) - that leaves us with 2.75 days - about right
  for a 3 day meeting

 [no objections on this schedule]

 LH - Re. the outreach item: we should talk/brainstorm to experts. We could telcon with them?

 LR - I think it's a good idea

  [No objections]

 [ AI - LR to follow up this idea - (due date before f2f). Email to Olivier, Janet & Judy to attempt
      to schedule expert/brainstorm meeting. ]

 KG - In this case I will need some names for security procedures

 LR - ok

 LH - I think this [proposed experts meeting] will be good to help us lay down some specific AIs and
      plans to advance us from our present conceptual ideas stage to the next more practical stage.

 [ There is discussion about where and when f2f attendees should meet on the first morning]

 KG - what are accommodation arrangements. Where do we meet 1st time on Monday?

 LH - How about the Marriot?

 KG - OK, lets meet in Marriot lobby, 8.00am

 [No objections]

 3. OpsGL leftovers

 [LH takes a vote on the proposed changes to checkpoints 6.1 and 4.6]

- CP 6.1: [no objections]

- CP 4.6: [no objections]

 [Both changes accepted by default]

- issue 59:

 LH - We should not try to have a global submitted test materials license - leave it
      to WGs to accept donations on terms that are consistent with other W3C principles.

 AT - In that case we should refer WGs to any documentation of the any such principles.

 LH - in OpsGL 5.3 - define licenses. It's peripheral but related to issue 59. Should we
      modify this? maybe should provide criteria & constraints.

 KG - sounds interesting. eg in XML schema WG just published a paper in which they explain what
      conditions a submitter must provide a license for.

  [ AI: LH to propose a rephrasing of CKPT 5.3  (assume due date Seattle f2f)]

4. Spec Guidelines

- 14.1 (also involves issue #99):

 [ The group debate the issue of whether the highlighted requirements are assertions
   or not ]

 LH - I think the correct reference in the agenda is [4]

 MS - [defines test assetion from specGL] These are expressed requirements. We cannot
       assume that requirements = assertions.

 LH - [Disagrees] - highlighted styled sections - I believe these are test assertions

 LR - A requirement is a checkpoint and the "to fulfill" part is the test assertion.

 LH - Consider checkpoint 1.1 - is this an assertion? If not how would it be expressed as
      an assertion?

 MS - The 'MUSTs' are requirements - we derive assertions from these so it is not
      correct to say that highlighted requirements are assertions.

 AT - Requirements explain what must be done to implement a specification. An
      assertion describes a (hypothetically) conformant product. So the language
      used is slightly different. E.g. a requirement says "The product MUST render
      the word 'tree' in green", an asertion would say "The conformant product
      renders the word 'tree' in green".

      The requirements in specGL could easily be made into assertions with a simple
      language mapping so it's a close call in this case but it may not always be
      so simple (in other specs).

 LH - So are you [AT] saying that they are assertions or not?

 AT - Technically no because the language used is slightly different. But it's very

 LR - Most people are trying to get a statement of what is required. We don't want
      to overburden them with attempting to be perfect about nit-picking details.
      It's analogous to security - things can be made so secure that no one can get
      their work done.
      I would prefer a 90% solution rather than something correct but unworkable.

 MS - Are we going to accept from other people that requirement == assertion?
      I don't agree that just pulling out the MUSTs meets the requirements.

 LH - It would help if we took some requirements examples from SpecGL and attempted
      to express them as assertions. This would identify any differences that we
      should condsider.


 [ AI: MS -  To provide assertion examples from SpecGL. prior to Seattle f2f ]

 KG - Do we link to examples? It may be useful to list those example in order to refine our


     [ AI: KG to provides links to XForms, SOAP and maybe others for ExTech }

- Checkpoint 1.4:

 MS - There is a reluctance of standards committees to include formal semantics instead of English
      language - on the pretext of difficulty of understanding. So if we have formal descriptions -
      we should give an example to help understanding

 LR - What do you mean by 'formal' - schema, BNF? other?. What were we trying to say? - the original
      intention seems to have been lost.

 LH - I recall that we wanted an example for everything 'example for every test assertion'
      Alex [Rousskov] objected to this - said it was excessive. Changed it to
      'at least 1 e.g. (ckpt1.3]' PLUS an example for everything that is hard to understand.

 LR - but we never came to conclusion of what a formal description was.

 [ Discussion to continue at next telcon, Wednesday 18 December ]


 LH - We will have another telcon on Wednesday to continue with the SpecGL issues.
      It will be 1 hour earlier than usual (10am EST)


Received on Monday, 23 December 2002 15:31:41 UTC