Re: Umbrella Specifications Discussion

Le mercredi 15 juin 2005 à 10:44 -0400, Karl Dubost a écrit :
> A profile is not, for example, a Primer, so is disqualified in my  
> model ;)

Yup :) This "primer" thingy seems a bit arbitrary to me, but again, I
can't judge without an answer to my earlier question.

> I have trouble with the question you are asking.  "special", what do  
> you mean?

My question is: what is the point of creating this concept? What is that
we want to say about "umbrella specifications" in Variability of
Specifications? For instance, we define the notion of "profiles",
"modules", "levels", etc because we need to describe how these
subdivisions interact with each other and how they affect the
conformance model of a specification.

We introduce the notion of "dimensions of variability" to explain a
higher level concept between the previous notions and to  communicate
better how adding more variation creates complexity and interoperability
issues.

But I can't tell at this point why we are introducing the concept of
"umbrella specification" since I don't know what we're trying to explain
about them or what we're trying to convey through them.

> In fact it's maybe the term which is almost not very well chosen:  
> Umbrella Specification.

(note that according the the minutes I quoted before, Patrick suggested
that term as the one used by the Java specifications, IIRC).

>  When I start discovering a technology by  
> looking at a document which defines it, there will be more than one  
> scenario.
> 
> * The Web technology is entirely defined, explained, presented in  
> only one document.
> or
> * The Web technology is defined in multiples documents that can be  
> parts of the technology, modules, profiles, etc.
> 
> In this second case, as an implementer/user, I want to be able to  
> know what is it about and how everything is tied together.
>
> Is it clearer?

Well, I see that you want this as a user, but how does this fit in
Variability in Specifications? What are the advanced topics regarding
variability and/or conformance that need to be made about these types of
specifications?

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