W3C home > Mailing lists > Public > www-qa-wg@w3.org > September 2004

Re: [ISSUE] Do quality control - Spec GL version 2004-08-30

From: Dominique HazaŽl-Massieux <dom@w3.org>
Date: Thu, 09 Sep 2004 17:12:14 +0200
To: karl Dubost <karl@w3.org>
Cc: www-qa-wg@w3.org
Message-Id: <1094742303.5545.240.camel@stratustier>
Le jeu 09/09/2004 ŗ 00:54, karl Dubost a ťcrit :
> Depending on the class of product, you can implement every guidelines 
> or not. For the ICS 34 to ICS 37, you can only reply if you are part of 
> the WG which designing the specification.

The idea of doing profiles for SpecGL is interesting, but it may be an
overkill considering that we wanted to keep our conformance model as
simple as possible.

Also, I think this in fact shows a problem that we had tried to solve in
the previous SpecGL but has resurfaced: our conformance requirements
(that is, the principles, but also probably the good practices) should
always refer to an end result, and not to an action.

Since we're using imperatives conf req in specLite (which I think is
good), it's easy to get confused about this, but typically:
"Identify who or what will implement the specification" can easily be
mapped into a particular end result in a spec (a list of classes of
products)
while "do a systematic and thorough review" can't.

I think it may be more consistent to "simply" downgrade the ones that
are not about end results into non-conformance requirements texts.

Dom
-- 
Dominique HazaŽl-Massieux - http://www.w3.org/People/Dom/
W3C/ERCIM
mailto:dom@w3.org


Received on Thursday, 9 September 2004 15:12:16 GMT

This archive was generated by hypermail 2.2.0 + w3c-0.30 : Thursday, 9 June 2005 12:13:18 GMT