Re: Cc/pp S&V conformance section

Dear Stephane

Thank you for the opportunity to comment on your conformance section and 
CC/PP document.  On behalf of the QA WG I have the following comments and 
suggestions.

Adding the classes of products for CC/PP helped to focus the 
requirements.  Developers in my organization found it helpful in allowing 
them to concentrate on the relevant areas of the specification and helped 
ensure that they didn't miss a requirement.

Comments on 5.4 CC/PP Consumer Conformance.

 >We defined a test suite in which we defined a set of features to test on
 >different consumer implementations to check the interoperability.
 >For each test, we are providing valid and invalid profiles to illustrate
 >it. However, we are not sure at all that an implementation passing all 
these tests
 >could be considered conformant to the spec and say so.  Indeed that just a
 >small set of tests to test interoperability and not to make an
 >extensive check of implementation capabilities and conformance. So we
 >have quite trouble defining the conformance criteria for the consumer.

Conformance tests are never exhaustive, so you can never test 
everything.  It is especially difficult to define tests for 
consumers.  That is why the Specification Guidelines checkpoint 11.3, 
requires a conformance disclaimer.

A difficult section to write, since the statements are somewhat circular, 
that is, they depend on one another.  The key is to describe what is 
required for a conformant consumer.  Keep in mind that it isn't always easy 
or possible to test for conformance, but that it should be clear to the 
developer what the implementation must be able to do.

1. What is meant by 'extracts appropriate information'.  Since 
'appropriate' is vague and subjective, can this be made more specific?

2. It may be clearer to be more explicit regarding a consumer's support (or 
non support) for Appendix B.  If I understand correctly, support for the 
RDF Schema is mandatory for validating consumers.

3. Editorial:  Change 'all' to 'any' in non-validating: A consumer is a 
CC/PP conformant non-validating consumer when it does not reject all 
non-conformant CC/PP documents.

4. Comments on 5.5.2 Well-formed.
Add to the list of information to be included in a claim, the name 
(identify) of the implementation to which the claim is being made. Also, a 
version, date, or other identifier should be included to uniquely identify 
the implementation.

regards
Lynne





At 09:29 AM 6/27/2003, boyera stephane wrote:
>Dear Lynne and QAWG,
>
>
>I'm sending you this mail on the behald of DIWG, following the comments
>you sent us wrt the LCWD of CC/PP Structure and Vocabularies.
>
>We eventually drafted a conformance section for the spec, i'm attaching
>it to this mail.
>Please review and comment it.
>
>However, we still have few questions we were not able to answer in the
>group: particularly we have problem with the definition of section 5.4
>of what is a conformant consumer:
>We defined a test suite in which we defined a set of features to test on
>different consumer implementations to check the interoperability.
>Foreach test, we are providing valid and invalid profiles to illustrate
>it.
>However, we are not sure at all that an implementation passing all these
>tests could be considered conformant to the spec and say so. Indeed that
>just a small set of tests to test interoperability and not to make an
>extensive check of implementation capabilities and conformance. So we
>have quite trouble defining the conformance criteria for the consumer.
>So we are waiting for advices on this point.
>
>Thanks a lot in advance for any help
>Stephane
>
>
>--
>Stephane Boyera         stephane@w3.org
>W3C                             +33 (0) 4 92 38 78 34
>BP 93                           fax: +33 (0) 4 92 38 78 22
>F-06902 Sophia Antipolis Cedex,
>France
>

Received on Monday, 14 July 2003 11:35:21 UTC