Re: Personal review of QAF

I see all my earlier comments on the QAF, particularly those in

http://lists.w3.org/Archives/Public/www-qa/2004Jan/0002

as adequately formally addressed by Patrick's presentation at the tech 
plenary.

http://www.w3.org/2004/Talks/tp2004-test-pc/

and I do not expect any further formal responses to my comments, or for 
detailed issue tracking etc, except where it is useful in your judgement.


I also withdraw my earlier formal objections, which are similarly adequately 
addressed.

Of course, I am still open to informal discussion, and hope that some of the 
detailed comments will still be useful to you. I am also open to fomal 
discussion, if, for example, you wish to have an audit trail  motivating 
changes to your charter. E.g. I would support changes to your charter 
dropping the AAA conformance commitment, in response to my comment that you 
did not appear to be meeting this.

It is currently unclear to me whether it is useful to give further feedback on 
the recently published Test WD, which I note includes some responses to 
issues I have raised. Would it be more sensible to wait for the next round of 
publications?

Further, given that WebOnt is only chartered through to May, and that you 
appear to have accepted the main comment from WebOnt

http://lists.w3.org/Archives/Public/www-qa/2004Jan/0001

[[[
We [...] believe that the changes
needed [...] will be large enough to force another Last
Call phase.
]]]

you may wish to consider whether to inform WebOnt that that part of the 
comment has been accepted, but more detailed consideration is unlikely to be 
complete before WebOnt ceases to exist.
Within WebOnt I would then argue for a similar carte blanche in our response 
to such a message.

I also found the other presentations at the tech plenary useful, and I am more 
aware that my position is at an extreme within the W3C and that a consensus 
framework may well emphasize the needs of conformance testing and may make 
some use of RFC 2119 keywords where the interoperability use cases are, for 
example, examination of CR results and constructions of software which 
implements a number of W3C recommendations. (However, I doubt that I would 
ever be convinced that such use cases require MUSTs rather than SHOULDs)

Good luck

Jeremy

Received on Monday, 8 March 2004 04:18:37 UTC