[sc34wg3] And yet another...

[Lars Marius Garshol]

* Jan Algermissen
| 
| Do you mean to say that the above goals are the goals the reference
| model part of the standard should meet? And if the reference model
| meets them, then it is a good model?

What I'm saying is that that was the goal I was interested in. If the
RM had met this goal I would have been interested in the RM, yes.
 
| Honestly, I think this is far too weak if the intention is that
| Topic Maps as a paradigm are to have any significant impact in the
| field of data modeling and information organisation.

That's your view.
 
| I am deeply convinced that the purpose (the overall goal) of the
| paradigm as such is the only source for real requirements that any
| structural model must meet.

As far as I'm concerned XTM 1.0 did this a long time ago, except it
didn't have a model. Now it does.
 
| IMHO, the 'flow of argumentation' must be:
| 
| "Topic Maps are a paradigm that is designed to solve this and that
| existing problem and in doing so, it will have this and that benefit
| for users of the paradigm.  The abstract information structure that
| underlies all topic maps must be designed to meet the requirements
| imposed be the overall goals of the paradigm. Therefore, this
| abstract information structure is as follows...."

As far as I'm concerned TMDM == Topic Maps, and this is how I've
evaluated TMDM. For me, the RM/FM/... discussion isn't on this level
at all.
 
| A precise statement on the goal of Topic Maps is the *only* source
| for evaluation of any proposed underlying model.

No, Jan, that's not a given. You may think so, but I don't. As far as
I'm concerned ISO already defined what topic maps are a long time ago
and what we're doing now is something else. 
 
| So, why don't we define this overall goal of Topic Maps *first*?

Good question. I've been asking about this for years, and you're
actually the first person to take up the question as far as I can
remember.
 
| Several things may happen:
| 
| 1) We all agree
| 
|    Fine, then we'd have a clear means of reasoning why your model,
|    Robert's or the current RM is best suited or why they are all
|    missing the critical points.
| 
| 2) We know what we think, but we totally disagree
| 
|    Less fine, but also ok. This will give us a basis for further
|    discussion and IMHO reveal the real source of the disagreements
|    over the last years.
| 
| 3) We do not know what to say
| 
|    Bad! Without knowing what the paradigm is to achieve, how can we
|    find an appropriate model? How can we ever argue, that what we
|    propose (RM included) is sufficient, if we do not even know what
|    it must be sufficient for?

We're on 3) now, as far as I know. We're most likely to move on to 2),
probably.
 
| So, would everyone mind to write down her or his understanding what
| this overall goal is? I have never read nor heard anything that
| would give me a clue. Honestly.
 
I'm with you on this, and I think I've done it. 
 
| <rant>
| There is surely more to Topic Maps then doing data integration
| (merging) on datasets that all share a common ontology
| (name,occurrence, class-instance, superclass subclass)?!?
| 
| If this is all, then a published relational schema would have done
| the job.
| </rant>

Frankly, I don't see that either TMDM or RM do any more than this.

-- 
Lars Marius Garshol, Ontopian         <URL: http://www.ontopia.net >
GSM: +47 98 21 55 50                  <URL: http://www.garshol.priv.no >