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

At 12:01 PM 8/18/2004 +0200, Dominique Hazaël-Massieux wrote:
>Le lun 16/08/2004 à 22:41, Karl Dubost a écrit :
> > Le 05 août 2004, à 15:04, Karl Dubost a écrit :
> > > Good Practice:
> > >     In the conformance clause, define how normative language is 
> expressed.
>
>So, trying to clarify what I was saying during Monday's teleconf:
>- I think we really mean how "conformance requirements are expressed";

I agree, that is a more accurate expression of the intent.

>I
>don't know what we would mean by "normative language", e.g. how does
>"normative language" relate to "normative content" [C2 does in fact uses
>the "conformance requirements" term rather than "normative language"]

In SpecHeavy, we had two checkpoints about what's normative:  the micro 
level (i.e., the expression style of individual conformance requirements, 
like sentences with RFC2119 keywords), and the macro level (sections, 
appendices, figures, examples, conformance categories, etc).  SpecLite 
attempts (or attempted) to preserve that distinction.

>- having reviewed quite a few W3C specifications, I know that I don't
>think it's a bug for anyone not to describe its conformance requirements
>style in the conformance section, i.e. I wouldn't ask anybody to change
>their specs if the information is already available in an obvious place,
>like a "Terminology" section; as such, I don't feel compelled to put
>this as a good practice, since I know I wouldn't in fact recommend it

I can think of at least two counter-examples.  One, SVG, has the typical 
verbiage about RFC2119 keywords, etc.  But there is *lots* of stuff that is 
normative that doesn't involve keywords.  Description of graphical effect 
of an element, for example -- it is ordinary descriptive prose, but really 
contains the detailed conformance requirements for graphical viewers.  Only 
"common sense" tells a reader that it's normative.  (The "Common Sense 
Conformance Model"!)

>- I agree that an option could be to relax the GP to allow linking from
>the conformance section rather than including in it; but I know as a
>spec author I would find that useless
>- I like that our new SpecGL is lite; creating a good practice for this
>looks too heavy for me
>
>I'm still of the opinion that this GP should be in C2, with a technique
>indicating to put it in the conformance section or in a terminology
>section

I still disagree.  But won't pursue the issue, as it's apparently a 
minority opinion.

-Lofton.

Received on Wednesday, 18 August 2004 16:05:21 UTC