[sc34wg3] Thoughts on the RM

Michel Biezunski sc34wg3@isotopicmaps.org
Fri, 25 Apr 2003 01:18:51 -0400

> It is not in 13250, in any way, shape or form. That is my
> point. What you or SRN claim (today) was in your heads at
> some point in the past is neither here nor there. We cannot
> build and promote a standard on that basis.

For the record, when Steve Newcomb, I and others 
invented the concept of topic, of association, of 
occurrence, etc., it also started in our heads. 
And a standard was built and promoted on that 
basis. What I am saying is that sometimes it 
happens that a standard is built on what is
in someone's head.

Actually, I don't see how anything, including a
standard, can be built without starting in someone's

> What we have to go on, today, is ISO 13250:2002 (and the
> XTM 1.0 Specification, as the "provisional explication" of
> the XTM DTD).
> Those documents provide the only authoritative basis for
> determining what topic maps are. They are insufficient for
> the purpose of implementing topic map-based systems. What
> we need to do is clarify them enough in order to be able
> to ensure that information owners do not waste their time
> when trying to base their solutions on ISO 13250.

I agree 100% with the need for clarification. The TMM 
model is aimed at making transparent what topic maps 
applications are supposed to be doing to be unambiguously
interchangeable by explicitating the elementary
aspects of what's going on. The TMM, once made 
understandable, can be extremeley useful to help
assessing the strengths and the weaknesses of existing 
TM-based applications. It's like a microscope. 
Now that we have one, if we make the effort to 
study how to use it, we might very well end up 
benefiting from it for the standard or non-standard
tm applications, existing or yet to come.

> >In any case, even if it was new, it's not a bad thing to
> >do new things from time to time. Everything rusts,
> >including standards. We can't hope preserving the model
> >we designed and pretty much finished in 1995 (except for
> >scopes) for the eternity.
> I agree. I am not talking about eternity. I am talking
> about now, and the next few years. Let's make the standard
> we have workable, quickly, so that we stand a chance of
> getting it accepted.

We need to get the nuts and bolts right to make sure 
that interchange really works. That's my understanding 
of the purpose of what everybody is doing now. The fact that we
have two different approaches to do it, the SAM and
the TMM, is helpful, because by confronting the two,
we can get a much deeper understanding of the issues
we are trying to solve, knowing that the final
goal here is to get stronger, fully interchangeable
tm applications to work, to serve the market which
demand them. This is the same objective that the
one you describe. 

I just don't understand why you say "quickly" since
it takes some time to assimilate the new concepts
that are being created, and we need to have a clear
idea of what we are doing, otherwise it's not a standard,
it's just a quick and dirty fix, which might break
as quickly as it has been done.

> At the same time, let's think towards the future. Let us
> by all means consider how we might generalize the ideas
> in 13250 even further, if there is a clear business case
> for doing so. But let us also recognize that work for what
> it is: a further generalization of Topic Maps, something
> new that will take time to get right. And let us avoid
> killing the baby we have by confusing the market into
> thinking it has to wait for the next baby if it wants to
> take advantage of the wonderful idea that is Topic Maps.

If that's what worries you, I think this is fixable.
I don't believe that the TMM as proposed endangers in
any way existing topic maps applications. It helps
understand what's going on under the hood, and we
need to uncover that to make sure that what we have
now is actually working as expected. The fact that
other things can be done as well with the TMM is not
in contradiction with strengthening what we have. 
TMM is not a replacement for SAM, it's being proposed
as a mechanism to generate SAM and similar models.

My vision is that we don't have to announce the
next baby, the next babies are already there, and
we can either adopt them under our umbrella, or
reject them by saying that they are not exactly
identical to the one we have. I think it's more
dangerous for the future of the standard to
reject the new babies than to make some room
to accommodate them, even if they don't look
exactely the way we expected. This is another
advantage I see how to use the TMM, once we
will have understood how it works, which I
hope we will in London.

Michel Biezunski
Coolheads Consulting
402 85th Street #5C
Brooklyn, New York 11209
Web  :http://www.coolheads.com
Voice: (718) 921-0901