Re: Is infinity equal to infinity?

Erik Runeson <erikr@iar.se> writes:
> When comparing floating-point numbers, should infinity (Inf) be
> concidered equal to infinity?


Mathematics: two sets A and B have the same size if there is a
bijection from elements of A to elements of B, (a table that uniquely
maps pairs of elements so that no element in either set remains
unmapped.)


According to this definition and if you assume Inf to be the number of
a denumerably infinite set A and Inf' that of a denumerably infinite
set B, you have Inf == Inf'.


That's an argument, but what you decide to do still depends on what
infinities you are dealing with. (Does is make sense to think of them
sizes of sets? If so, are the sets denumerable, i.e. you give an
algorithm that returns any argument in a finite period of time?)


Don't know if this helps?




  -Matthias
--
Max-Planck-Institut für Informatik | Deutsches Forschungszentrum für KI
fis@mpi-sb.mpg.de | fischman@dfki.de
http://www.mpi-sb.mpg.de/~fis |
--

Return to the comp.compilers page.
Search the comp.compilers archives again.