New paths for subject identity? RE: [topicmapmail] Making ontologies : RDF vs TM

[Lars Marius Garshol]

* Bernard Vatant
| 
| Ditto. Along those lines, I've been questioning lately here in an
| exchange with Lars Marius the "absolute" relevancy of subject
| indicators, IOW what are the relevant contexts in which subject
| indicators can/should be used for identification?

I think I should expand a bit on what I wrote earlier. I think the
standard is right to say that merging on subject indicators is
compulsory, and I think the reason this is right is that it makes the
interpretation of topic map data predictable. If you *don't* know that
the recipient will be doing subject indicator merges then they may be
looking at that that is wholly different from what you were looking at
when you sent it, and consequently queries that worked for you may not
work for them, and data that validated successfully for you may not
for them, etc

In addition, when storing a reference to a topic in this scenario, you
cannot use the subject indicator if the topic has one (which is a
tragedy, since that's the only firm identifier you've got), since you
can't be certain that things will merge correctly later, if someone
turns off that merging.

And so on and so forth. There are a lot of reasons why I think this is
a crucial piece of behaviour in a TM system, but I won't go into all
of them.

Of course, this doesn't mean that everybody should be using subject
indicators for all identity-related information. What it means is that
the standard provides you with predictable, if inflexible, behaviour,
and that if you want different behaviour you can use a different
modelling approach, for example an occurrence. You can then define
your own rules for using these for identification, and everyone gets
to have their cake and feast on it, too.

| Yes. If identification needs a context, then how do you *identify
| the context*? 

If subject indicators are absolute, that gives you a way to bottom out
when you've made a pragmatic decision that you are as far down in the
recursive well as you want to be.

| If we follow those tracks, yes, identity can only be established on
| a pragmatic basis. 

True.

-- 
Lars Marius Garshol, Ontopian         <URL: http://www.ontopia.net >
GSM: +47 98 21 55 50                  <URL: http://www.garshol.priv.no >