[sc34wg3] Documenting merging rules in TMDM .. and unmerging too

[Michel Biezunski]

On Wed, 2004-03-17 at 05:01, Lars Marius Garshol wrote:
> * Bernard Vatant
> | 
> | And to answer the final (BV?), yes, my view is that the standard
> | should define the declarative semantics (subject identity), and be
> | agnostic about the operational ones (merging, mapping,
> | whatever). Although it could provide non-normative best practices
> | about the latter.
> 
> Uh, so your opinion is that we should take the document currently
> published as CD and entirely change our view of what it is and what it
> does, and then rewrite it to support that?
> 
> Or do I misunderstand you?

Good to see that you are open to the idea, Lars. It would certainly be
worthwhile to reconsider. Making a clear distinction such as the one
proposed by Bernard between the levels we are addressing would
strenghten the different parts rather than weaken them. The difference
between the declarative parts and the operational parts is at the
heart of both SGML and XML and that is I think why this family of
standards have been so successful. If we could succeed in doing
something similar for the semantic part, wouldn't that be great?

Michel
 
-- 
==================================
Michel Biezunski
Coolheads Consulting
402 85th Street #5C
Brooklyn NY 11209
Email: mb@coolheads.com
Web  : http://www.coolheads.com
Voice: (718) 921-0901
============v======================