Re: First-Order Programming Theories

Walter G. Wilson <WALTER@ISI.EDU>
Message-id: <9307061332.AA18641@aragorn.watson.ibm.com>
To: cg@cs.umn.edu
Cc: interlingua@ISI.EDU, sowa@turing.pacss.binghamton.edu
Subject: Re: First-Order Programming Theories
In-reply-to: Your message of Sun, 04 Jul 93 18:02:29 EDT.
             <9307042202.AA04657@turing.pacss.binghamton.edu>
Date: Tue, 06 Jul 93 09:32:44 -0500
From: Walter G. Wilson <WALTER@ISI.EDU>



>>In Part IV, the authors develop their version of "time logic", which
>>is a many-sorted FOL with time as one of the sorts.  They show that
>>time logic is more general than (i.e. a refinement of) the dynamic
>>logics and temporal logics that they discussed in Parts II and III.
>>They do that by embedding both dynamic logic and temporal logic in
>>time logic.

One of the problems I have with time as a sort or as an argument of
predicates (eg as when used as a method to represent changing state)
is its relativity -- not only theoretically, but practically: it is
difficult to get identical timestamps on distributed systems. Give me an
epsilon and I can eventually get the systems withing that epsilon,
but it will take work, (and then will be subject to clock drift or exceed
the mean-time-to-failure if the epsilon is too small...)

Does he address these problems of the meaning of time as percieved by
machines in a distributed system?

Walter