Repeated convolution of a continous uniformdistribution



I have a fix for abs_integrate, but I've run into problems with some
limits;
for example

 (%i6) limit(signum(x-t),t,inf);
 (%o6) limit(signum(x-t),t,inf)

Of course, (%o6) could be -1. It would be great if limit could be extended
to
handle this case.

Barton

-----maxima-bounces at math.utexas.edu wrote: -----


>I?understand?the?bug.?It's?not?easy?to?fix--it's?a?mathematical
>error,?not?a?programming?bug?:(