Re: Discrete Logic

Carlos Gershenson (carlos@jlagunez.iquimica.unam.mx)
Thu, 21 May 1998 19:41:15 +0200 (MET DST)

Its a philosobhical problem, with real-world applications.

I agree with you that it is now impossible to solve problem to have
infinite number of sets.

But, who knows? We can evaluate an integral form minus infinite to
intinite.

What I am trying to say, is that the more sets you consider, the more
realisticly your problem will be modeled. Maybe someday we could use sets
as real numbers, and play limits with them...

Greetings,
____________________________________
Carlos Gershenson
carlos@jlagunez.iquimica.unam.mx
http://132.248.11.4/~carlos/


On Mon, 18 May 1998, WSiler wrote:

> In a message dated 98-05-01 08:24:16 EDT, you write:
>
> > But if we want the degree of truthness of something, we could do, as in
> > fuzzy logic, to use fuzzy sets. And every set would give a different
> > thruth, and overlapping them, we would get THE truth. But there's a
> > problem. In the universe, there are infinite number of sets for anything
> > we want to truthify. What can we do?
> >
> Is your question a philosophical questions, concerned with what we call
> things, or is it a practical question, concerned with how we approach the
> problem of real-world reasoning? If it is a practical question, then perhaps
> we have something to talk about; but if it is a philosophical question, I'm
> afraid I don't understand the question very well. Perhaps it is somewhere in
> between these Boolean possibilities!
>
> >From a practical viewpoint, we reason in limited domains. If we wish to
> consider the possible truth of a number of statements such as X is good, X is
> neutral, X is bad, we construct a fuzzy set of such statements. We then have
> some data pertaining to X and to our value system, and we need some fuzzy
> rules which relate these inputs to the truth of the various possible
> statements about X, and away we go.
>
> If we wish to consider an unlimited set of possibilities, in which the
> possibilities are not numeric values, then I consider this to be an ill-formed
> problem, and can be of no help.
>
> William Siler
>