On 12/7/06, miguel lopez <miguel39123 at hotmail.com> wrote:
> This is an example of multiple integral in maxima:
>
> Integrate(x^3*y^3,[x,a,b],[y,c,d]);
I'm in favor of extending the integral notation in Maxima, although
I don't see a need for the initial capital letter.
Every now and then I think Maxima should recognize integrals
like \int_{\Omega} f d(\mu) i.e. let the region be more general than
an interval, and let the measure be more general than Lebesgue
measure, and while we're at it, recognize the named function f
(or a lambda expression) instead of requiring an expression.
I think it would be OK to figure out a representation first, and
think about computations later.
My own interest in this is Bayesian inference, in which you'd like
to set up everything in terms of general integrals, and then handle
special cases after plugging in problem-specific stuff.
E.g. some of the integrals turn into summations if the measure
is a discrete measure, others turn into 1-d ordinary integrals,
others 2-d or 3-d, etc. To make that work, Maxima needn't know
much about general integrals, except, say, Fubini's theorem and
stuff like that.
FWIW
Robert Dodier