Newsgroups: comp.object.logic,sci.logic,sci.philosophy.tech
Path: cantaloupe.srv.cs.cmu.edu!rochester!miller
From: miller@cs.rochester.edu (Brad Miller)
Subject: Re: logic of change
In-Reply-To: jand@cs.kuleuven.ac.be's message of 31 Aug 1994 09:14:13 GMT
Message-ID: <MILLER.94Sep2170347@wolverine.cs.rochester.edu>
Lines: 34
Sender: miller@cs.rochester.edu (Brad Miller)
Organization: University of Rochester
References: <341hl5$4c7@idefix.CS.kuleuven.ac.be>
Date: 02 Sep 1994 21:03:47 GMT
Xref: glinda.oz.cs.cmu.edu comp.object.logic:236 sci.logic:8109 sci.philosophy.tech:15722

>>>>> "Jan" == Jan Dockx <jand@cs.kuleuven.ac.be> writes:

    Jan> I need a logic, preferably based on FOL, but not necesseraly, that provides a system for describing the fact
    Jan> that the result of the evaluation of a clause depends on the time of evaluation. In other words, the set of
    Jan> facts and axioms can change over time.  What I need more specifically is a description of "dynamic sets". Given
    Jan> a set S, and t1, t2 elements of T (time), I want to be able to speak about S(t1) and S(t2), which are regular,
    Jan> classic mathematical sets. It may be that S(t1) = S(t2), but if t1 <> t2 it is possible that S(t1) <>
    Jan> S(t2). Maybe S(t2) has a few elements more or less. So, it is possible that for some x, "x element of S" is
    Jan> true when the clause is evaluated at t1, and false when it is evaluated at t2.

    Jan> Anyone? Pointers to "dynamic logic" (b[PROGRAM]e), temporal logic (event calculus, modal tense logic (
    Jan> P(P(F(clause))) ), situation calculus, Allen's theory of time (periods)) are not needed. They don't do what I
    Jan> need.

Allen's relations don't work by themselves, but you can attatch them to the sets.

I.e. [Holds [p] [t1]] would mean p is true during interval t1.

Similarly you can do set membership (modulo equality issues), i.e. you can
deal with a set-membership predicate, but not have objects o1 and o2 equal
during different intervals; objects are either equal or not, there is no temporal
dependancy.

See Hans Koomen's Thesis, available as a TR from here (contact peg@cs.rochester.edu),
and the TEMPOS system, part of RHET, available for FTP from 
ftp.cs.rochester.edu:/pub/knowledge-tools. One of the README files you can get in 
that directory will point you to various TRs that you can ftp as well. (but doesn't
include his thesis).

If you can't ftp, everything is also available on the CMU archive CD-ROMs, 
from ptf@cfcl.com (I'm not affiliated).



