Use of maxima in 'statistical engineering'

Hi Vishal,

For the record, I don't know anything about Maxima at NIST.

It seems you & I have similar interests in statistical computing.
An idea which I have pursued off & on for a long time is to state
a problem in probabilistic inference, try to get an exact result
and then fall back on numerical methods if that fails.
I can send you some references to stuff I wrote if you like.

About computing convolutions, one method is to discretize the
pdf's and compute the convolution as the inverse Fourier transform
of the product of Fourier transforms. Maybe that's applicable in
your case.

I'm interested in symbolic computations with functions defined
piecewise, as those appear in practical problems very often.
I have made some attempts at that but I haven't gotten far.
I think others are working on it too.

If there are some papers you've written, I'd be interested to
take a look at them.

Thanks for your interest in Maxima,

Robert Dodier