OASIS Mailing List ArchivesView the OASIS mailing list archive below
or browse/search using MarkMail.


Help: OASIS Mailing Lists Help | MarkMail Help

tm-pubsubj message

[Date Prev] | [Thread Prev] | [Thread Next] | [Date Next] -- [Date Index] | [Thread Index] | [List Home]

Subject: RE: [tm-pubsubj] Contradicting a PSI


> | 'A is not identical to B' means 'it exists some assertion true for B that
> | is false for A'.
> | This is the simplest way to say two subjects are different : you can say
> | something which is true for one and false for the other.
> Yeuk. Please, let's not go *there*! As SRN is fond of saying:
> You should never confuse topic maps with the truth!
> Neither should you restrict topic maps to being non-contradictory.

Well, I agree that maybe I went too far. But OTOH, we should make provision for
consistency when needed, which means have rules allowing to keep my TM consistent if I
want them to be so, that I can relax if I don't care about consistency.

> The CIA might have a PSI for Osama that states he's a terrorist. Do
> I have to agree with that assertion in order to use the corresponding
> PSID with integrity? I sincerely hope not.

I'm not as sure about that point than you are. Why would one use CIA's PSI at all if one
does not agree with CIA, IOW, if one does not trust CIA, or disagrees with their
definition of "terrorist" and/or the fact that Osama is one instance of this class.
Suppose Al Qaida has its own PSIs set, with completely different definitions for
"terrorist" and "Osama".
If your topic map uses both and want to merge CIA's stuff with Al Qaida stuff, you are
responsible of the mess involved, although the two original publishers are internally

> The point is to know whether we are talking about the same thing...
> not whether we have the same opinions about it.

Making sure we are talking about the same thing implies we agree on a minimal set of
properties (ka assertions, or opinions) about it. And we are here about the difficult
distinction about subject "constituting" properties and "other" properties, and IMO this
is the same distinction the RM makes between SIDPs and "other properties". So I would
rephrase the above as following.

'A is not identical to B' means 'it exists some *identifying property* true for B that
A does not share'.

And it's the responsibility of the PSI publisher to be crystal clear about which
properties are identifying the subject, and if you don't agree with them, then don't use
this PSI: you are speaking about something else. So, in your example, all depends if CIA's
PSI for Osama declares his being a terrorist as a constitutive property, or not.

This principle may help to answer the question : what kind of information should I put in
a PSI?


PS : See also my quick message on SC34 about merging rules.

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

[Date Prev] | [Thread Prev] | [Thread Next] | [Date Next] -- [Date Index] | [Thread Index] | [List Home]