[sc34wg3] Re: [topicmaps-comment] RE: OASIS vs W3C

H. Holger Rath sc34wg3@isotopicmaps.org
Tue, 25 Sep 2001 20:19:09 +0200

Hi Sam,

I am cross-posting this to the SC34 WG3 mailing list for the same
reason I cross-posted my other email: this discussion is very 
valueable to find the proper reqs. for the core data model.
Sam, I hope you don't mind.

You wrote:
> Hi--
> > I doubt TMPM4 can do what RDFS already can and what TMCL will do.
> Possibly; possibly -- cardinality, for instance; we'll have to see.

TMPM4 constraints associations. But much more can be constrained:
topics, occurrences, scopes, etc.

> (Quite frankly, as I listened in the ISO meeting to the (very fine)
> list of requirements that Lars put together, I was struck by how many
> of them were already considered and handled by TMPM4.)

I see a simple assoc template, but nothing more.
> > More discussions on this should take place in the ISO mailing list.
> Possibly; possibly not. I'm not trying to start a flame war here, or
> break anybody's rice bowl (to use yet another cliche) -- but I think
> that at least the ideas in TMPM4 deserve consideration, attention, and
> circulation. 

Yes, of course and all discussions between LarsM and SteveN show that
we have progress because "people talk" to each other and bring the
arguments on the table.

> As the possesor of a doctorate, I'm sure you will agree
> with me that open, unsuppressed discussion of all ideas is as good (and
> necessary) for the topic map community as it is in the academic
> environment

I agree 200%. But part of this discussion is raising (new) issues
(e.g., my issue of taking templating out of the core data model)
and looking at new developments (e.g., TMCL).
> TMPM4 has, at least, been through several drafts and been implemented
> several times -- as opposed to TMCL, which at this point is only a
> straw man. 

Scientific progress lives from this fact. There are experiences with
existing stuff and there are ideas how things could be done in the
future. And the discussion leads to new results and new discussions. Etc.

The ISO committee is not a scientific group but a team of people
with certain interests. The common interests of this group are:

- make TMs a success
- define standards necessary to achive this success

Science works for centuries in the described infinite loop but ISO work
has to finished at some point in time. This means that procedures known
from product development (reqs analyses, use cases, functional reqs,
design, implementation, testing, corrections, approval) have to be applied.
Otherwise no defined and controlled outcome will be produced.

> To use yet a third cliche, to compare the two at this point
> is really to compare "apples and oranges."

It is my goal of this discussion to make the core data model and
TMCL apples and oranges - because they are completely different things,
even if TMCL requires the core data model. And to achive this goal
all constraining stuff has to be taken out of the core data model. QED.