[sc34wg3] N0393 Topic Maps Model

[Michel Biezunski]

The TMM represents in my opinion a huge
progress over everything we had before
(PMTM4), RM, etc.

Thanks to Steve N., Sam H., Jan A., and 
Patrick D., who have spent an obviously 
tremendous amount of time to achieve 
this result.

I believe that we are very close to the
intent of what Topic Maps mean since the
beginning of this effort and the first
versions of the standard. In particular,
the fact that topics are described as 
the unique place in electronic space where
all the information about a given subject
is to be located is exactly what topic
maps are about.

I am going to work more on this text
because I believe that there are still
some ways to express things that can
be improved, but I need time to do that.

For the time being, I think the interesting
effect of this model is that it enables
a diversity of applications that can still
interoperate at a fundamental level. 
The SAM model and the models which are
implicit in XTM and HyTm (presumably very
similar) become particular applications 
of Topic Maps and are -- hopefully -- fully
interchangeable. It also means that some
of the issues that have been discussed
regarding the SAM applications can
be resolved in one way rather than another 
even if it doesn't satisfy all applications,
because depending of the context, 
it's possible to design an application 
that precisely fits a set of needs which 
can be slightly different from another, 
and these different families of applications 
can still be interchanged as topic maps. 
One example (but there are plenty of those) 
is the absence of subject constitutor in HyTm. 
It's now OK to have a topic map application 
making a difference betwen subject constitutor 
and subject indicator and another taking only 
account subject indicators. There is no need
to "fix" HyTm to make it possible.

It also seems to me that because of the level
of genericity, there are other applications out
there (relational databases among plenty of others)
that become interchangeable as topic maps while having
no notion of an XML syntax to interchange them.
I suppose we will need to explicitly state
how this is doable, but from what we have now it
looks very possible. 

Thanks again to the authors of this wonderful 
piece of work. I look forward to seeing you all 
in London to discuss more about how
this affects the structure of the future Topic
Maps standard.

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



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