Re: [SpecGL Draft] A.1 GP In the conformance clause, define how normative language is expressed.

At 04:41 PM 8/16/2004 -0400, Karl Dubost wrote:
>Resumé of the discussion.
>Le 05 août 2004, à 15:04, Karl Dubost a écrit :
>>Good Practice:
>>         In the conformance clause, define how normative language is 
>> expressed.
>This Good Practice, that has been written, imposes that:
>The normative language is given in the conformance clause. (which means 
>that the usual terminology section has to move to the conformance clause 
>for any specification which would like to conform.)
>* Keeping this good practice means:
>         - Removing the terminology section of SpecGL to move it to the 
> Conformance Clause

I disagree.  I support the idea that it could stay where it is, as long as 
there is a link to it from the Conformance Clause.

We argued this at length during LC of SpecGL -- e.g., do the test 
assertions actually have to be in the spec?  or can they be linked?  We 
decided "linked" would suffice to meet the intent of "included".  I think 
the same applies here.  The key concept is:  you should be able to find all 
important conformance information by starting in the Conf. Clause.

>         - In reviews asking to other people to do it as well (though it's 
> a good practice, not a principle, so not mandatory)

I think we are going to find that this is a problem for other GPs as 
well.  I guess we have now decided that Principle is like MUST, and Good 
Practice is like RECOMMENDED.

>         - Usually the conformance section is toward the end of the 
> documents, which means have the terminology information only at the end, 
> except if they do the effort to read the conformance section first.

Then link it (terminology) from CC, if you want it to appear earlier.  I 
don't see the problem here.

>* Removing the practice means:
>         - Rewrite a bit the principles in section C.2
>                 """Principle: Use a consistent style for
>                 conformance requirements and explain
>                 how to distinguish them"""
>         like adding a technique to explain how to do it and how to create 
> a Terminology section.

I believe that the GP should stay in A.1, and be harmonized with what's in 
C.2 (or perhaps removed from C.2, and in C.2 place a link or reference to 
its occurrence in A.1).

>         - Leave the terminology section at the top.

Not an issue, linking suffices.

>         - Adding a link in our conformance clause saying where the 
> terminology is defined.

Ummm... I'm afraid I'm getting lost here.

>Another Issue:
>         The way we have written the principles don't use the RFC 2119 
> wording, which is fine for me, as long as we define their mandatory 
> nature in the conformance clause and we explain what a MUST, SHOULD means 
> in the rest of the prose.

I agree, the RFC2119 terminology section says what these words mean when 
they occur (*if* they occur [?]).  Does not mean (or should not mean) that 
we are using the 2119 keyword to express our individual conformance 
requirements (Principles).

>         Though after checking it's not really issue... because it seems 
> we are not using RFC2119 keywords for setting implementation requirements 
> :)))))



Received on Tuesday, 17 August 2004 22:06:08 UTC