By the way, this was me, not Franklin (who doesn't exist). I am using a new
laptop (my hard disk etc crashed Friday) and the mail account set up for me
had my name wrong. (01)
Leo (02)
-----Original Message-----
From: ontolog-forum-bounces@xxxxxxxxxxxxxxxx
[mailto:ontolog-forum-bounces@xxxxxxxxxxxxxxxx] On Behalf Of Franklin L
Obrst
Sent: Sunday, February 22, 2004 10:42 AM
To: cassidy@xxxxxxxxx; '[ontolog-forum] '
Subject: RE: [ontolog-forum] Frames vs Logic again (03)
This is true of all frame systems that I know of, i.e., slots. Slots are not
really predicates or relations (which are for binary relations,
bidirectional as Adam says), but instead are "directed". If the underlying
OKBC model of Protégé supported n-ary "relations", it would still be the
case that one argument is focused on, since frame systems are
object-centered, and OKBC is a frame system. Yes, I would add inverse slots
for all of these, if you use Protégé, indeed a duplication. In OO-like
systems (like frame-based systems) slots are typically used for
"attributes", and the inverse is usually not called out. (04)
More purely logic-based systems like KIF don't have this problem.
Unfortunately, frame systems are kind of a relic from earlier, pre-logical
AI, and in fact were created as alternatives to logic, not a good idea. (05)
Description logics (which seek less expressive logics in general than FOL
[1], to enable more tractable automated reasoning) formalize and extend
frame systems. In DLs, you'll find a different notion, i.e., roles, which
muddy the situation a bit more, since roles are rather like slots. A good
tutorial on Description Logics is Franconi's at:
http://www.inf.unibz.it/~franconi/dl/course/. (06)
For a comparision of the various formalisms, see [2], off the above
tutorial. (07)
[1] But some DLs are actually more expressive than FOL, e.g., if they allow
for transitive closure (of roles) or fixpoints.
[2] D. Nardi, U. Sattler, D. Calvanese, R. Molitor. Relationships with
other formalisms. In the Description Logic Handbook, edited by F. Baader, D.
Calvanese, D.L. McGuinness, D. Nardi, P.F. Patel-Schneider, Cambridge
University Press, 2002, pages 142-183.
http://www.inf.unibz.it/~franconi/dl/course/dlhb/dlhb-04.pdf. (08)
Leo (09)
-----Original Message-----
From: ontolog-forum-bounces@xxxxxxxxxxxxxxxx
[mailto:ontolog-forum-bounces@xxxxxxxxxxxxxxxx] On Behalf Of Patrick Cassidy
Sent: Saturday, February 21, 2004 11:58 AM
To: [ontolog-forum]
Subject: Re: [ontolog-forum] Frames vs Logic again (010)
Yes, hasName and several other concepts that neither add to nor detract
from the logical content were added to take advantage of the visual
representation capabilities of Protege. If they bother anyone, they
can be sequestered in a separate module and left out of versions
that don't get imported into Protege.
It is also possible in Protege to create windows with lists
of propositions where a given concept is in position 2 or 3 or whatever,
just as the SUMO browser and Sevcenko's browser have done. But at
this point I haven't implemented that, since that is already covered
by the other browsers. It is not a built-in function in Protege.
To cover that issue, I have only added a few inverse relations
thus far for those cases where the presence of a
concept as argument 2 in the SUMO relation can be of significance in
understanding a concept. If it turns out that viewing relations with
classes as argument 2 is of more general interest, I can create
a separate pane for the Class window that displays such relations. (011)
Pat
================ (012)
Adam Pease wrote: (013)
> Pat,
> I noticed in your file of suggested additions the following:
>
> --------------------------
> (instance hasName BinaryRelation)
> (domain hasName 1 Entity)
> (domain hasName 2 SymbolicString)
> (inverse hasName names)
> (documentation hasName "hasName relates an instance of an entity to a
> string of linguistic characters used to reference the entity in
> linguistic communication. This is the inverse of the SUMO relation
> 'names', added to allow more flexible representation in Protege. The
> hasName relation is not a necessary relation since not every entity is
> named by a SymbolicString and not every SymbolicString is the label for
> an entity.")
> ----------------------------
>
> This is entirely redundant with the existing SUMO relation 'names'. The
> only reason one would want such a definition, is, as you note, to
> overcome the limitations of a frame system. A frame system is oriented
> to inspection and reasoning on the first argument, so if one looks at
> the frame for 'SymbolicString' one won't see the slot 'names'. It will
> only be visible when one is looking at the frame for 'Entity'.
>
> If you feel the need for this inverse of 'names' a case could be made
> that every binary relation must also have an inverse, thus doubling
> (uneccessarily) the number of relations. Of course, this also doesn't
> solve the problem that all ternary and higher order relations (of which
> there are a number in SUMO) will still be invisible in Protege or any
> other frame system.
>
> This is a good example of why Protege is a bad choice for a formal
> ontology expressed in logic.
>
> Adam
>
> _________________________________________________________________
> Message Archives: http://ontolog.cim3.net/forum/ontolog-forum/
> Subscribe/Unsubscribe/Config:
> http://ontolog.cim3.net/mailman/listinfo/ontolog-forum/
> Shared Files: http://ontolog.cim3.net/file/
> Community Wiki: http://ontolog.cim3.net/wiki/ To Post:
> mailto:ontolog-forum@xxxxxxxxxxxxxxxx
>
> (014)
--
=============================================
Patrick Cassidy (015)
MICRA, Inc. || (908) 561-3416
735 Belvidere Ave. || (908) 668-5252 (if no answer)
Plainfield, NJ 07062-2054 || (908) 668-5904 (fax) (016)
internet: cassidy@xxxxxxxxx
============================================= (017)
_________________________________________________________________
Message Archives: http://ontolog.cim3.net/forum/ontolog-forum/
Subscribe/Unsubscribe/Config:
http://ontolog.cim3.net/mailman/listinfo/ontolog-forum/
Shared Files: http://ontolog.cim3.net/file/
Community Wiki: http://ontolog.cim3.net/wiki/
To Post: mailto:ontolog-forum@xxxxxxxxxxxxxxxx (018)
_________________________________________________________________
Message Archives: http://ontolog.cim3.net/forum/ontolog-forum/
Subscribe/Unsubscribe/Config:
http://ontolog.cim3.net/mailman/listinfo/ontolog-forum/
Shared Files: http://ontolog.cim3.net/file/
Community Wiki: http://ontolog.cim3.net/wiki/
To Post: mailto:ontolog-forum@xxxxxxxxxxxxxxxx (019)
_________________________________________________________________
Message Archives: http://ontolog.cim3.net/forum/ontolog-forum/
Subscribe/Unsubscribe/Config:
http://ontolog.cim3.net/mailman/listinfo/ontolog-forum/
Shared Files: http://ontolog.cim3.net/file/
Community Wiki: http://ontolog.cim3.net/wiki/
To Post: mailto:ontolog-forum@xxxxxxxxxxxxxxxx (020)
|