John(B)
The following is copied from a response to John(s) on the AESIG list
about FOL with respect to our needs to integrate architectures:
1) First order isn't good enough - architectures are modal,
non-monotonic and deontic - in other words, all the stuff that keeps
logicians up at night but people use all the time. While CL may be able
to go beyond FOL and there may be some ways to encode some of these in
FOL+, the semantic set of modeling languages is essentially open and so
must be the ecosystem.
2) It is too hard - we want the ecosystem to be friendly to well defined
languages as well as those that are not. We want to be able to import
knowledge that may not be grounded or well formed, we may or may not
ground it later. To require that every concept that is brought into the
ecosystem be formally grounded is too high a bar and would prevent
adoption.
3) It is fragile and does not deal with inconsistency well. (01)
So this will probably get me in deep water on this list, but it seems
like there is an over-emphasis on FOL. We want to represent and make
sense of the knowledge people have and the way the want to express it.
The concerns of what is computable a FOL reasoner is not the first
priority. There are certainly advantages to the computable subset of a
model and the strong semantics of FOL, but that is to provide a certain
set of capabilities and to "ground" those concepts that can be grounded
in a well established logic. But, other logics have value as well and
are part of the overall picture. (02)
-Cory Casanave (03)
-----Original Message-----
From: sio-dev-bounces@xxxxxxxxxxxxxxxx
[mailto:sio-dev-bounces@xxxxxxxxxxxxxxxx] On Behalf Of John Bateman
Sent: Thursday, April 15, 2010 2:05 PM
To: [sio-dev] discussion
Cc: AESIG
Subject: Re: [sio-dev] Sharing and Integrating Ontologies (04)
David: (05)
> What is it that CASL can do that Common Logic is not capable of doing? (06)
nothing I would hope: they should both be FOL! :-) (07)
But, more seriously, the focus here is
not on CASL or CL, or CASL vs. CL: that
violates entirely the principles of heterogeneity that we follow.
The focus at present is on the *structural layer* (HetCASL) that
provides a formally specified (and implemented in HETS) mechanism
for gluing together specifications in any logic that is known to
the system. Adding CL into HETS should then automatically provide
CL with access to the reasoning tools connected to HETS as well as
providing a well-founding structuring layer on top of CL: that is
the aim at least. (08)
From our ontology work we are convinced that effective ontology
design requires such structuring mechanisms but they are not
provided in CL (or OWL or ...). After adding CL into HETS,
whether one then uses CASL or CL to write ontologies
should not really matter. (09)
> What needs to be done for a logic to be a "comparable logic" ? (010)
I meant for the logics *to be* comparable, rather than doing
something to them: e.g., CASL and CL will
be comparable in that they are FOL-ish. Clearly linking CASL
and a DL profile of OWL is going to loose stuff and so they are
not directly comparable. Actually it is still an open question I think
just how comparable CASL and CL are in the last resort since
there are some differences in the finer details. (011)
Best,
John B. (012)
P.S. In case it is useful, though, here is some more further
background on CASL. (013)
CASL is integrated via HETS with a wide range of reasoning
tools (unlike CL) and has been used for many years in larger-scale
algebraic software specifications. (014)
From the website: (015)
"The Casl design effort started in September 1995, as a common effort of (016)
the Compass Working Group and IFIP WG1.3 (Foundations of System
Specification). An initial design was proposed in May 1997 (with a
language summary, abstract syntax, formal semantics, but no agreed
concrete syntax) and tentatively approved by IFIP WG1.3." (017)
"the design was finalized in April 1998, with a complete draft language
summary available, including concrete syntax. Casl version 1.0 was
released in October 1998, and Casl version 1.0.1 was officially approved (018)
by IFIP WG1.3 in April 2001. The present version 1.0.2 is documented in
the Casl Reference Manual, and illustrated in the Casl User Manual (both (019)
published by Springer in LNCS in 2004)." (020)
It is now a de facto standard for algebraic specifications with
extensive tool support and so is the language we tend to use for
our own ontology work. (021)
_________________________________________________________________
Msg Archives: http://ontolog.cim3.net/forum/sio-dev/
Join Community:
http://ontolog.cim3.net/cgi-bin/wiki.pl?WikiHomePage#nid1J
Subscribe/Config: http://ontolog.cim3.net/mailman/listinfo/sio-dev/
Unsubscribe: mailto:sio-dev-leave@xxxxxxxxxxxxxxxx
Community Shared Files: http://ontolog.cim3.net/file/work/SIO/
Community Wiki:
http://ontolog.cim3.net/cgi-bin/wiki.pl?SharingIntegratingOntologies (022)
_________________________________________________________________
Msg Archives: http://ontolog.cim3.net/forum/sio-dev/
Join Community: http://ontolog.cim3.net/cgi-bin/wiki.pl?WikiHomePage#nid1J
Subscribe/Config: http://ontolog.cim3.net/mailman/listinfo/sio-dev/
Unsubscribe: mailto:sio-dev-leave@xxxxxxxxxxxxxxxx
Community Shared Files: http://ontolog.cim3.net/file/work/SIO/
Community Wiki:
http://ontolog.cim3.net/cgi-bin/wiki.pl?SharingIntegratingOntologies (023)
|