- From: Dominique Hazaël-Massieux <dom@w3.org>
- Date: 12 Aug 2002 17:00:27 +0200
- To: David Marston/Cambridge/IBM <david_marston@us.ibm.com>
- Cc: www-qa@w3.org
Le lun 12/08/2002 à 16:40, David Marston/Cambridge/IBM a écrit : > Dominique writes: > >- I wonder if the GL about levels brings anything useful, since levels > >don't really exert any influence on conformance or implementation. > > It was a surprise to read that! I think the influence of levels is > clear when you consider the verbiage constraints in Checkpoint 11.2, > at least after the word "levels" in GL 11 is replaced. 11.2 would, > in my interpretation, require that a spec that uses levels (in the GL > 7 sense), to prescribe level-aware verbiage for conformance claims. > Example: All claims of conformance to this spec should state the level > of implementation, [levels enumerated here], which is claimed. > > Thus, one implementation could claim to fully implement the Xblah spec > at Level 2, while another implementation claims conformance to Xblah > Level 3. Well, that's one more time the issue of separating the concepts of specification and technologies: you don't conform to the DOM (the generic technology Document Object Model), but you conform to the specification "DOM Level 1" or the specification "DOM Level 2". The hierarchical relation betweens two levels - e.g. the fact that DOM Level 1 has all its features included in DOM Level 2 - doesn't imply anything obvious on the conformance claim, or does it? Since the GL 11.2 [1] already implies that you must specify the "specification name" which includes the Level (any counter-example), I'm not sure we need to give further details on this. I may miss something obvious, of course. Dom 1. http://www.w3.org/QA/WG/2002/08/qaframe-spec-0804.html#Ck-claim-wording -- Dominique Hazaël-Massieux - http://www.w3.org/People/Dom/ W3C/INRIA mailto:dom@w3.org
Received on Monday, 12 August 2002 11:00:29 UTC