On Sun, Jul 7, 2013 at 5:50 PM, Barton Willis <willisb at unk.edu> wrote:
> > Does this mean Maxima cannot do DiracDelta() calculus correctly no
> matter what I try?
>
> No it doesn't mean that. If x * delta(x) / x --> delta(x) is a deal
> breaker, you'll need to change the general simplifier.
> Of course modifying the general simplifier might introduce new bugs, break
> lots of code, or slow calculations. So it's
> an activity that requires a great deal of care.
>
> Arguably x/x --> 1 is an incorrect simplification; nevertheless it's
> possible to do useful work with Maxima, by the way
>
Agreed, and something like what Fateman has been talking about for years
(packaging assumptions like x#0 with the result) could help.
But I fear that if we tried to do things like that correctly across the
board, we'd end up with a system that was much harder to use for relatively
simple applications, especially since we do such a poor job of simplifying
conditionals.
Re x*delta(x)/x, what is the *disciplined* way of handling that? Why do
you simplify (x*delta(x)) first? Why doesn't it associate as
x*(delta(x)/x), which presumably is undefined at zero? (or is it?)
-s
> --Barton
>
> _______________________________________________
> Maxima mailing list
> Maxima at math.utexas.edu
> http://www.math.utexas.edu/mailman/listinfo/maxima
>
>