Re: SUO: should a standard specify...
Jean-Luc,
Thanks for your comments.
Some replies (and much silent snipping) below.
Jim
------------------------------------------------------------------------
JF > How to get the relevant ontological primitives and assertions from the
> ontology itself?
J-L: In order to "understand" the content of any ontology you have to have
a description of it's syntax and semantics, even for "hand-carving".
JF: Acknowledged.
JF> What I want to suggest is that the assumption of a common language for
> expressing different ontologies may not be realistic, at least not at the
> present time. (See, for example, the discussion below of "thing" in SUMO,
> Cyc, and ISO15926-2.)
J-L: You are assuming because the word "thing" is used with different meanings
by different people/specifications it is impossible to find a "true" meaning.
JF: No, I am not assuming this, at least I didn't think I was in what
I was trying to say.
J-L: You are right in that the meaning is not and *cannot* be the same, but you
are wrong in assuming that to make sense of interoperability between such
different contexts you have to reconcile the meanings based on syntactical
identity or even similarity of the symbols used. A SUMO thing, a Cyc thing
and an ISO15926-2 thing need *not* have the same meaning.
JF: Again, I wasn't trying to suggest that the way to get at meanings was
through syntax alone (and apologies if it came across that way). What I
wanted to suggest was that while as a human I can more or less successfully
grope my way through different (English) words that might have similar
meanings, without some mechanized way to search for potentially similar
terms, my human-based procedure quickly becomes ineffective. So I was
wondering if there could be additional guidance provided either within
the ontologies themselves or in their specifications that could help
automate my groping.
JF> What I want further to suggest is that maybe what is needed is not so much
> an agreed-upon common language, but an agreed-upon method to identify in
> a given ontology what are the primitive entities, relations, axioms, and
> definitions, and how these primitives can be used to extract
> domain-specific subontologies. Perhaps such an agreed-upon method should be
> part of the ontology's specification. Such a method would certainly help in
> the examples below.
JL: Such an "agreed-upon method" is obviously dependent of the *target*
"common logical language" which is yet to be defined. This might be CL but
I see no reason for a more easyly reached (and sensible...) agreement
on CL than on SUO.
JF: I'm not sure the "agreed-upon method" necessarily depends on
the target "common logical language." I leave open the possibility
that it may, but I don't see that it necessarily does.
JL: Then you expect that ontologies "who happen to find themselves on a street
corner" will be reconciled by some process *without* some common semantic
grounds, however small this may be (I do think it can be *very* small).
You can as well ask for the ontology broker to have "psychic" capabilities
that will be a *real breakthrough* in AI!!!
JF: No. I expect that they do need some common ground, and I am willing
to grant that the ground be in some sense "semantic."
What I am suggesting is that this common ground could perhaps be part
of what an ontology standard specifies -- a type of protocol
for how ontologies go about interacting with each other.
> I want to find those components of the different ontologies that bear
> on, say, naive spatial relations. But even before I try to extract those
> components, I'll probably need to understand something very general: how do
> Cyc, SUMO, and ISO1596-2 treat the notion of "thing" or "entity"?
J-L: Exactly, but this has to be given a *formal* definition.
JF: Acknowledged.
JF: > So, I looked in each of these ontologies for "thing" and "entity".
> What follows are some results, which don't pretend to be exhaustive.
[snip]
JF: > How are people or programs to deal with such variability?
J-L: It is not so much the variability than the *lack of content* of
these so-called definitions that is the problem.
An ID is an ID, period. Any "meaning" it takes come from the rules
that are specified to operate on and between the IDs.
JF: I agree that the meaning won't come from, say, an ID. And I agree
that the lack of content could in some cases be an issue. But even
if lack of content is one of the issues here, it is not the main one
that I am trying to address, which is: how to make available to programs
the ontological guts (including language elements, axioms, model theory,
"content", etc.) of ontologies, so that users or programs stand a chance
of being able to judge whether two particular ontologies might be
able to be used together?
JF: > (In a sense, perhaps, what I'm looking for is a sort of generic ontology API
> that would allow me to gather the relevant ontological primitives for a
> particular domain. Ideally, such an API would allow programs to automatically
> assess different ontologies for compatability, based on only a small set of
> seed primitives.)
J-L: That's about what I propose, it has yet to be defined.
JF: I will be curious to hear what others think.
JF: > As far as the SUO effort is concerned, I suggest that we consider whether
> the identification of and access to certain ontological primitives, as well
> as the means to use those primitives for extracting domain-specific
> subontologies, should be part of the standard.
J-L: I beg to differ. Why the heck is everybody so keen to slap yet another
layer of API/specs/primitives onto an existing structure each time a
specific need arise?
JF: Indeed. And others in the group may well agree with you. I am not eager
to foist on the group considerations that the group feels are best
dealt with elsewhere. I just wanted to pose the possibility that the
group may want to deal with these considerations, since they seem
to be of practical importance in determining the extent to which two
ontologies might be able to interact.
J-L: Cannot you deal with the already present capabilities built in the
structure? There must be some, otherwise there would not even be any structure to be seen.
JF: I can. And I may well end up cobbling together my examples by hand, using
the existing structure from one or more of the works I mentioned.
J-F: > ------------------------------------------------------------------------------
> SUMMARY
[snip]
> * The main difficulties seem to be: (1) how to identify the relevant
> ontological primitives (e.g., terms, definitions, axioms) used by an ontology;
> and (2) how to use these primitives to extract a particular domain-related
> portion of the ontology.
J-L: Why would you want to do (2)?
If the ontology is there why not just use it?
JF: Overhead and targeted domains. I may want to acquire and evaluate what
10 different ontologies say about naive spatial reasoning, without having
to crunch through the other 95% of what the ontologies say about the world.
(I realize that it may be difficult to extract "what an ontology says"
about a particular domain and that one needs a way of determining how to
delimit a domain.)
JF: > * Being able to identify, extract, and manipulate these primitives _seems_
> key to being able to assess whether two ontologies can interoperate
> (in virtually any plausible sense of that word). But it may indeed not be
> appropriate to suppose that one could determine whether two ontologies
> can interoperate based on accessing the ontological primitives and
> assertions used by the ontologies to talk about a particular domain.
> To those who hold this view, can you please elaborate?
J-L: I did partly already, are you sure you read carefully enough and
extracted all the info available?
JF: If I understand you correctly, you are asking whether I extracted
from SUMO, Cyc, and ISO15296 all the info relevant to "thing" or "entity"?
If that is what you are asking, then my answer is "No, I am not sure
I extracted all the relevant info." And indeed finding a way to talk
about such relevant info is part of what I am after.
JF: > * Currently, as shown by the examples of SUMO, Cyc, and ISO15926, different
> ontologies have different ways of representing these primitives, and in many
> cases it is difficult to determine from the ontology itself, how these
> primitives are identified and can be made available. Currently, these
> determinations must be done manually (I think) and in an ad hoc way.
> Perhaps the SUO standard can address how to accomplish make such
> determinations programatically.
J-L: Exactly. This requires a "common" description convention for both the
syntax and semantics that could be applied to *all* such formalisms.
And this is DIFFERENT FROM YET ANOTHER COMPETING formalism within the
same domain. It does NOT have to be used for anything else than describing
the other formalisms, just like EBNF versus a specified language.
JF: This is interesting. Again, I will be curious to hear what others
have to say.
[snip]
Thanks,
Jim