Re: "the graph" WAS RE: [xtm-wg] Reification of topic map constructs

["Steven R. Newcomb" <srn@coolheads.com>]

[Sam Hunting:]
> If in fact there are three layers presented, of markup, processing
> model, conceptual model, with three notations to master (XML, graphs,
> UML) that may be too much for the "Desperate Topic Map Hacker" to
> absorb, and lengthy to boot. I will gladly trade partial understanding
> by early adopters for ubiquity.

> Therefore, explicating what Steve calls "the graph" is best left in a
> supplementary (though authoritative, ie XTM-branded) monograph. We
> don't want to end up building a boat in the basement -- though perfect
> in concept and construction, it is too big for the basement stairs, and
> hence cannot actually sail anywhere. 

> [daniel]
> > For me, the topic map graph *is* the topic map.

> Well, as was famously said in another context, it depends on what the
> meaning of "is" is. To me the topic map "*is*" the markup (my mission,
> if you will, is about long-lived content, and "data outlives
> programs"). About this we are not likely to agree, and need not even
> argue, as long as we agree (which we do) that "the graph" is the result
> of a topic map markup as processed (presumably by a topic map engine). 

What I've been calling the "topic map graph" is, in fact, the
*significance* of the XTM-conforming portion(s) of XML documents.
It doesn't depend on implementations.

I think we should *try* to avoid using the term "topic map graph"
because it's really too (and confusingly too) close to seeming to
prescribe how XTM processors should be implemented.  Instead, we
should *try*, if we possibly can, to use circumlocutions that remain
agnostic about implementation methodologies.  I'm not really sure that
we can, but we should try.

For example, I don't think we should use phraseology such as, "<topic>
elements become objects that conform to the "topic" object class
defined in the conceptual model," even if we believe that to be the
only sensible way to implement a topic map processor.  We shouldn't
give the impression that XTM constrains implementations to such an
extent that it prescribes what's an object and what's not an object in
any conforming implementation.

Instead, how about, "The information conveyed via <topic> elements
corresponds to the concept of 'topic' in the conceptual model."  I'm
not proposing this sentence, which in any case is inaccurate.  What
I'm proposing is "corresponds to the concept of" as a possibly
appropriate circumlocution.  No, on second thought, I'm not even
proposing *that*.  What I'm saying is that we have to decide, now, on
the kinds of words we're going to use in order to establish, in XTM
1.0, the correspondences between the conceptual and syntactic models.
I think we have to make those connections explicit *somehow*, even if
only at such a high level that some necessary guidance for
implementers is lacking.  If we don't say in *any* way what
corresponds to what, I think we would richly deserve the criticism
that XTM 1.0 consists of totally incomprehensible fragments whose
interrelationships are only alleged to be significant.

Given the space limitations we've imposed on ourselves, I don't think
we can do more than establish correspondences at a very high level.
But I would argue that we can't do less, either.

I'm wondering whether paired diagrams would help to make the
correspondences easier to convey, easier to understand, and less
space-consuming.

-Steve

--
Steven R. Newcomb, Consultant
srn@coolheads.com

voice: +1 972 359 8160
fax:   +1 972 359 0270

405 Flagler Court
Allen, Texas 75013-2821 USA

-------------------------- eGroups Sponsor -------------------------~-~>
eGroups eLerts
It's Easy. It's Fun. Best of All, it's Free!
http://click.egroups.com/1/9698/1/_/337252/_/974237594/
---------------------------------------------------------------------_->

To Post a message, send it to:   xtm-wg@eGroups.com

To Unsubscribe, send a blank message to: xtm-wg-unsubscribe@eGroups.com