FINAL minutes QA WG Telcon 20030903

QA Working Group Teleconference
Wednesday, 03-September-2003
Scribe: Dimitris Dimitriadis

(PC) Patrick Curran (Sun Microsystems)
(dd) Dimitris Dimitriadis (Ontologicon)
(KD) Karl Dubost (W3C, WG co-chair)
(DH) Dominique HazaŽl-Massieux (W3C)
(LH) Lofton Henderson (CGMO - WG co-chair)
(VV) Vanitha Venkatraman (Sun Microsystems)
(SM) Sandra Martinez (NIST)

(LR) Lynne Rosenthal (NIST - IG co-chair)
(MS) Mark Skall (NIST)
(AT) Andrew Thackrah (Open Group)

(PF) Peter Fawcett (RealNetworks)
(KG) Kirill Gavrylyuk (Microsoft)

(DM) David Marston

Summary of New Action Items:
AI-20030903-1 Karl Dubost reply to Dan Connolly that we will work very 
closely with other WGs to improve the framework 20030905
AI-20030903-2 Patrick Curran to redraft two sentences about interaction 
between QA and WAI and send to Lofton 20030903
AI-20030903-3 Lofton Henderson to send an updated version of reference 
4 20030907
AI-20030903-4 Patrick Curran to draft a discussion of assertions for 
the concepts section of TestGL 20030915
AI-20030903-5 Patrick Curran  publish issues list for TestGL (not 
necessarily final form) 20031003

Previous Telcon Minutes:


1.) roll call 11am EDT, membership

2.) Any routine business
††††††††[- overdue Action Items ]

3.) Status of OpsGL transition to CR
††††††††- Last minute change request (DC): y/n? [1]

(lh) One comment came from Dan Connolly. He suggested OpsGL was too 
dry, even with revised intro. QA framework intro starts off with 
overview of this doc, I linked from our first bullet. I think DC 
proposes there is a new section at the beginning. Practically I have no 
time to do this.
(pc) I agree with DC that it is rather dull. Do not think we can 
significally change, but have to live with it.
(lh) Most similar documents are like this. We can prove the readability.
(pc) Real world stories are a good idea, in case we have them they 
belong to the ExTech documents.
(lh) Anyone supports making this change?


(lh) Then we'll let it die for now. As a courtesy, should we send a 
message to Dan saying that his point is well taken but that we cannot 
do justice to this right now?

(kd)  we should reply to Dan Connolly that we will work very closely 
with other WGs to improve the framework and the editorial part.

4.) Any SpecGL topics
††††††††- DoC position on WAI issues [2]
††††††††- draft DoC [3]

Looking at the reference 2 on the issue about interaction between QA 
and WAI. We have endorsed that we need to have a consensus position to 
point at, looking at reference 2 to see if that is a good position. 
Patrick suggested that point 7 was unnecessary, Lofton agrees.
(pc) Tone down 8 as well.
(lh) Agree
(lh) any other comments?

5.) Progression of "Intro"
††††††††- approval of issue resolutions [4]
††††††††- Intro draft "W3C WG Note"

(lh) Caught up this week, written resolutions on all issues pertaining 
to the introduction.
(lh) Will send an updated version of the document by the end of this 
week (reference 4)

6.) TestGL topics (continued)
††††††††- continue at #2 in [5]
††††††††- previous refs:† [6a], [6b], [6c], [6d]

(pc) Item 2 is quick. We talked 2 weeks ago about overlap. We overlook 
licensing. In Crete we talked about categories of test materials. 
Possibility to license it. We said that the discussion of licenses 
should go into TestGL. It is however specifically adressed in OpsGL. I 
agree that this is the right place for it. Proposing to leave it in 

(pc) Back to discussion started a couple of weeks ago on assertions and 
metadata. I want to try to follow up on the discussion we started about 
cases where we believe assertions can be derived from the spec, or test 
cases derived from the spec. This is a possibility but the guidelines 
that says you should create assertions and so on clearly do not apply. 
Alternative would be to take all metadata and assert it with test cases 
rather than assertions.

(pc) Alternative interpretation is to list the particular expression 
from the grammar and say it is an assertion. What do we meen when we 
talk of an assertion?
(sm) There are two definitions of assertion. We should aim at 
consistency. In the SpecGL definition we have a richer definition.
(lh) We have a resolution that the glossary definition should be 
changed to match the SpecGL one.
(pc) sounds like you're saying there are cases where there are no 
assertions at all
(sm) there are cases where we have atomic only assertions
(pc) anyone who believes there is a one to one correspondece between 
assertions and test cases is wrong, it is easy to imagine one assertion 
being associated with up to a dozen test cases.
(pc) for me an assertion is fundamental. I would argue that you cannot 
have a test case without an assertion. That assertion is either present 
in the spec or you can point to it and explicitly explain how you 
derive it
(sm) What then is a test case description?
(sm) will try to come up with something in writing
(lh) before we move on, the whole discussion makes me think on whether 
we need to describe different types of testing. Seems to me we spent 
much time lately talking about test assertions and conformance 
requirements, maybe we should describe differences in a concept section.
(pc) it really needs explaination
(pc) If you cannot point directly to an assertion in spec, the test is 
invalid. you have to be able to point directly to an assertion, not 
just a section in a specification for example.
[dm joins]
(dd) adds to the burden of the author
(pc) language should be precise enough, the concept of assertions 
themselves does not put a bigger burden on the authors. test developers 
may have a different view
(dm) would not shy away from saying specs should do more
(pc) most specs do not tag specifically.
(lh) we should draft something
(pc) I'll put something together
(lh) also, we need to be better at keeping TestGL issues
(pc) have done so informally
(pc) AI publish issues list for TestGL (not necessarily final form), 
one month

7.) Adjourn
adjourned at 12.08

8.) Overflow (12-12:30): available.

Received on Friday, 12 September 2003 08:40:22 UTC