Kal, Tom, Murray, and al
Unfortunately I have not the bandwidth to dive deeply into this cheerful
debate, I just
quickly browsed through the thread trying to figure what it is about. Well, I
guess the
title of the thread is quite good and says it all, it's about Subject
Identifiers
*Metadata*.
Why do not we assume simply that, as any other set of data, a topic map can and
should
have metadata, and that it might be a good idea not to mix them up with the
data level,
which implies maybe forget about topic maps to express them, or at least have
different
layers linked by a specific mechanism which has nothing to do with TM, to avoid
what Tom
calls very accurately "semantic noise" caused mainly by usual recursivity traps.
Just FYI, in Mondeca ITM we had always used such proper metadata to capture
e.g. the
"conditions of creation" of any object in the TM (topic, association, role,
name,
occurrence, identifier ...), like creation date, creator, quality, source ...
all that
sort of DC stuff. And those metadata are represented in a completely specific,
internal,
proprietary way, since they deal with the history and management of the system
itself,
which is orthogonal to the information the system manages in the TM, and has
barely any
meaning outside ofthe system.
So, in ITM, it's quite easy to know which user attached a given PSI to a given
topic, and
when it was done. Now if you are interested by the history of the system by
itself, and
want to express it in a topic map, I don't think it would need too much
metaphysics. All
objects created are very well identified. Take the logs of the system, and
figure a TM
with that if you like.
Another thing, more philosophical. I see in the debate here and there coming
back
expressions like "real world" and "exist", sometimes both in the same sentence.
Hmmm ... I
remember my main contribution (maybe the only significant one) to XTM
specification was to
insist to have those words striken out. Subjects are subjects of conversation.
Existence
of the conversation is hard to deny. As for the existence of subjects outside of
conversation, well, I'm personally completely agnostic ... but, hush, there are
some
platonicians around :))
Bernard Vatant
Senior Consultant
Knowledge Engineering
Mondeca - www.mondeca.com
bernard.vatant-eA+GF8qqnh1BDgjK7y7TUQ@xxxxxxxxxxxxxxxx
> -----Message d'origine-----
> De : topicmapmail-admin-Zo64W7twoUFWk0Htik3J/w@xxxxxxxxxxxxxxxx
> [mailto:topicmapmail-admin-Zo64W7twoUFWk0Htik3J/w@xxxxxxxxxxxxxxxx]De la part
> de Thomas B. Passin
> Envoye : mardi 4 mai 2004 06:23
> A : topicmapmail-Zo64W7twoUFWk0Htik3J/w@xxxxxxxxxxxxxxxx
> Objet : Re: [topicmapmail] Subject Identifiers metadata
>
>
> Murray Altheim wrote:
>
> > Tom,
> >
> > I've been watching you and Kal go around on this one for awhile
> > now, and neither of you seem to be getting any closer to an
> > understanding. Not having been present during the conversations
> > I believe Kal is alluding to, you seem to be having trouble
> > catching that what he's talking about has nothing to do with
> > computers or computer systems, that the concept of "the thing
> > we hang topics off of" (what Steve Newcomb and Eliot Kimber at
> > one time were calling a "binding point") *doesn't* exist in the
> > same way that a mathematical point doesn't exist (except as a
> > point),
>
> Murray -
>
> As you have probably seen by now, I have gotten to the same point, since
> we chewed our way through the various other possibilities.
>
> > which is the reason that there's no way to reify it. In
> > the conceptual model for topic maps we didn't include the concept
> > of binding point, but it assuredly both exists and doesn't exist,
> > just as the thing that can't be reified because its only purpose
> > is to *be* a binding point for topic characteristics that exist
> > but don't exist, except in a vacuum. This is in essense the "topic
> > identity" and should not be reified because it does not exist. If
> > it existed we could put a cherry on top of it and eat it.
> >
>
> I'm afraid I always have trouble with this kind of language. I'm sure
> that it is just me and how my mind works. I need to have it be either
> simpler or more concrete. To the extent that I understand it, I'm not
> sure I agree that there is such a "binding point", either in the
> Platonic cave (which I am not at all sure about either), or in the way
> our minds work (which I **certainly** am not sure about).
>
> I will go further. Considering that I really don't know what a
> "concept" is, nor a "relationship", I have quite enough to deal with
> just mapping the computer constructs to these mental ones that I
> experience (and that I imagine others experience). I don't need any
> "binding points" blocking what little understanding I do have. If
> someone wants to use the term "binding point" as a shorthand for "the
> subject that a topic is mapped to", I would be happy with that. Up
> until today, I thought that was what it was supposed to be. Anything
> else, I am just going to ignore it, for to me it would just be semantic
> noise that would interfere with my meager understanding of a very murky
> area.
>
> I'm glad it conveys something useful to you, but it's noise to me.
>
> As for a mathematical point, whether or not there is such a thing in the
> physical world, I am clear that there is such a concept that has been
> floating around as a subject of discourse for millennia. *That* concept
> I am comfortable with representing by a topic.
>
> Be that as it may, I claim that if it is whatever you have been talking
> about in your previous paragraphs, if it is a reasonably well-defined
> concept you can create a topic for it (maybe with the help of a PSI that
> contains the text you just posted). Let's not get into what "reasonably
> well-defined" means, OK? Please, please?
>
> OTOH, if it's not a reasonably clear notion, or if there is nothing to
> be said, well, the point is moot, is it not?
>
> Cheers,
>
> Tom P
> _______________________________________________
> topicmapmail mailing list
> topicmapmail-Zo64W7twoUFWk0Htik3J/w@xxxxxxxxxxxxxxxx
> http://www.infoloom.com/mailman/listinfo/topicmapmail
>
|
