[topicmapmail] SIDPs and Identification Process [Was : Identities and names ... ]

[Bernard Vatant]

Lars Marius

> * Bernard Vatant
> |
> | Note that XTM distinction between <subjectIndicatorRef> and
> | <baseName> [...] look now quite short-sighted in this perspective.
>
> Are you sure you mean this? The semantics of the two are after all
> different, as are the merging behaviours. I'd be interested to hear
> how you would prefer to handle this.

I mean exactly what I say.

First, intended semantics were maybe different, but at the light of current
state of reflection about identification, they draw a line of distinction
in a somehow artificial place. Long debate about TNC until its eventual
relaxing (should I say death?) has clearly shown that identity issues and
naming issues were not so clearly distinct.

Second, the way I would prefer now to handle it now would be to specify in
TM a primitive notion of "identifying property" - or identifying
characteristic, like nonAmbiguous Property (in DAML) or inverseFunctional
Property (in OWL), and in fact I would gladly strike *anything about name*
altogether, letting open to applications the ability to declare any type
(subclass) of identifying property, name, URI or otherwise. This has been
expressed for a while - in some mathematical way in
http://www.mondeca.com/english/published-doc/hypergraph/hg4tm.htm

And again, I think that this is basically what SRN is up to with SIDPs.

> | - A subject is assigned characteristics or properties - including
> | names - through assertions.
> | - An identifying property (SIDP) is an assertion about a subject
> | that, in a given context, following a given protocol, makes for
> | subject identification (that is, ability to declare that two
> | subjects are identical or not). SIDPs can be assignment of names, or
> | other assertions.
> | - A semantic application or language has to clearly declare what are
> | the properties it uses as SIDPs.
> | - Semantic applications or languages are interoperable if they are
> | able to match their respective identifying properties, so that two
> | subjects identical for one of them are also considered identical by
> | the other(s).
>
> I think we all agree on this conceptually, although we use different
> terminology to say it.

I don't know whom you include in "we all", and I would be curious to see an
intensional or extensional definition of it, before celebrating loudly :)

> The question is: now that we've agreed on this,
> what should we do.

There again, the answer depends on whom is "we" but clearly, AFAIK, it
should be fairly broader that TM community, including at least RDF
community, of course the work on Published Subjects, and certainly others
to be defined.

> I'd be especially interested to hear what you
> think, Bernard. It's all very well to put forward an analysis, but the
> real question is what to do in practice.

Sure enough. This is indeed intended to make something happen, not just to
simmer more conceptual soup. But the first step is to define the "we all"
above, including finding out the relevant organizational framework.

> Do we want to change any part of the current standards proposals, or
> do we not?

If you are speaking about what is currently going on in ISO like SAM the
answer is clearly no. This work is about to wrap up things at the level of
a certain type of application using specific identifying properties.

> Does this relate to them at all?

Of course it does, in the sense that it intends to leverage the crucial
work that has been achieved in TM community about identification (even if I
consider the current specification not to reflect the latest stage of
reflection about it - and that is natural for a specification) as a basis
for extended work. The essential question is to know if this extended work
is leading Topic Maps to another stage of development, or if TM have to
stick to their original model (including the specific SIDPs defined in ISO
13250 and SAM) and therefore appear as a particular instance of a wider set
of semantic applications, in which a common extended definition of
identification would be used for interoperability. Those are technically
and politically completely different routes.

As far as I understand, original thinking of the supporters of the TMM was
to have that game played in TM land framework (this is the first route).
But I'm not sure they are not beyond this stage of reflection now. I think
you would certainly agree that the second one is better. That means as wide
as possible extension of the "we all". So the story is now quite political
: who is wanting to play in that space is the first issue to settle.

Bernard Vatant
Senior Consultant
Knowledge Engineering
Mondeca - www.mondeca.com
bernard.vatant@mondeca.com