[sc34wg3] Illustrating SIDPs

[Patrick Durusau]

Robert,

Robert Barta wrote:
> On Fri, May 07, 2004 at 12:36:25PM -0400, Patrick Durusau wrote:
> 
<snip>

> 
>>Third, the TMRM does not constrain the formalism, or any other
>>aspect of writing, specifying or implementing a TMA.
> 
> 
> That is not so obvious.
> 
> If the TMRM specifies directly _what_ kind of statements about _what_
> kind of concepts have to be done in the process of 'disclosing', then
> this is a constraint on the formalism.
> 
> I think, we hit here a crucial point...
> 

Note the difference in our statements:

PLD: "...the TMRM does not constrain _the_ formalism..."

RB: "...is a constraint _on_ the formalism."

In other words, the TMRM does not constrain the formalism you choose to 
use for a TMA, but yes, it does constrain the concepts that the 
formalism would be addressing.

I did not think that particularly remarkable, particularly in light of 
Steve Pepper's oft repeated comments about what is or is not a topic 
map. The TMRM is not designed to nor do its authors have any desire to 
constraint what should or should not be said in designing architectural 
blueprints for a building. I am sure there are standards for that sort 
of thing and personally I would have no idea where to start.


> 
>>What the TMRM is trying to do is create an inventory and checklist
>>for a disclosure statement that you can use to construct a TMA
>>however you like.
> 
> 
> ...namely that - for me - the TMRM is like a half-constructed
> bridge. It is firmly grounded on one side of the river, stretches
> halfway over the water and towards the other side it only reaches out
> with some planks. Fitting in the end of the other side seems to be
> difficult without a means to configure this.
> 

Then you must consider the rules for an EBNF grammar a half-constructed 
bridge as well. By themselves, the rules for an EBNF grammar is an 
intellectual curiousity. With other components, however, those rules 
become quite useful.

BTW, "half-constructed" and "some planks" is really not enough detail 
for me to provide a more satisfying reply.

> As I understood the objective of this exercise, though, is (a) "to
> capture the essence of the TM paradigm" and (b) "to provide a mapping
> into the application level".
> 
> While I would have chosen a different path for (a) in the first place,
> (b) is not satisfying for me, seeing this through the eyes of a
> developer/systems integrator.
> 

We have a difference of opinion on:

 > (b) "to provide a mapping into the application level"

I agree that is a very necessary step

I disagree that such a mapping is of necessity the responsibility of the 
TMRM.

The rules for an EBNF grammar does not specify how such a grammar would 
be used at the "application  level," but that does not detract from 
their utility.

Given that you want to keep some unspecified "formalism" out of the 
TMRM, I am curious why you now seem to want a mapping directly to the 
application level in the TMRM? A serious question since you have yet to 
say what "capture the essence of the TM paradigm" means, at least as you 
use the term.


> I could not - as a vendor now - provide a clear depiction how my
> application works. Its like having a lot of atoms on the table, but
> only alchemistic means to build 'stuff'.
>

Without a TMA (an unfortunate term, I prefer disclosure since it does 
not overload "application"), there is no way to provide the depiction 
you seek.

However, without rules for that depiction, not only would your depiction 
not make sense to anyone else, it could well miss something that is in 
the "essence of the TM paradigm."

> 
>>Granted that the language of the TMRM can and should be simplified but 
>>identity is a complex question and to some degree the TMRM reflects that 
>>fact.
> 
> 
> I share that observation. Maybe identity can be factored out of TMRM?
> 

No. If you are looking for the essence of the TM paradigm, subject 
identity is the keystone you were looking for.

Looking at the TMDM for a moment, note that it goes to great lengths to 
define when two topic information items will be deemed to be equal and 
what happens when such is the case.

Aside to Lars: Even though we disagree on the details (a lot of those), 
wouldn't you say that subject identity and what happens when topics, 
associations or occurrences that represent subjects with same identity 
is the keystone to the TM paradigm?

(Not asking about our relative positions on the TMRM or TMDM, just as a 
general principle, however we think we should get there, or in some 
opinions may already be there.)

Hope you are having a great day!

Patrick

-- 
Patrick Durusau
Director of Research and Development
Society of Biblical Literature
Patrick.Durusau@sbl-site.org
Chair, V1 - Text Processing: Office and Publishing Systems Interface
Co-Editor, ISO 13250, Topic Maps -- Reference Model

Topic Maps: Human, not artificial, intelligence at work!