I half-suspect that these are problems that arise out of the GEMM
testing machinery which cause difficulty. Asking for a solution
would be a cheap and easy way to get "valid" answers. But I'm
a curmudgeon by nature.

I'm impressed with the true talent displayed by certain
mathematicians in solving these challenges. I wish I knew
how to automate such talent and intuition.

But the point of this group, at least for me, is related to the
"symbolic" nature of the topics. I'm less interested in the
answer than in how it can be automated. It's the old request
that the student should "show your work". How I can get your
answer automatically is more important than the answer.

The posted challenge would be more useful to symbolic
software (note that I have a vested interest in Axiom) if the
original challenge showed the failure of some symbolic
algorithms. The posted result would be more useful if it updated
an existing algorithm to handle the challenge. Such posts would
advance the state of the art.

So I don't have a problem with the idea of "challenges" but I do
feel that they are like "sunday crossword puzzle games", which
are entertaining but not adding to the state of the art.

Manuel Bronstein published a book on the algorithms for
integration. Given a challenge problem, show where his algorithm
fails and publish a change so that it can handle the challenge.

Where are the algorithms?

Tim Daly
Axiom Lead Developer

I would bet a good part is taken using
from Prudnikov et al.

d...@axiom-developer.org wrote:
I half-suspect that these are problems that arise out of the GEMM
testing machinery which cause difficulty. Asking for a solution
would be a cheap and easy way to get "valid" answers. But I'm
a curmudgeon by nature.

I'm impressed with the true talent displayed by certain
mathematicians in solving these challenges. I wish I knew
how to automate such talent and intuition.

But the point of this group, at least for me, is related to the
"symbolic" nature of the topics. I'm less interested in the
answer than in how it can be automated. It's the old request
that the student should "show your work". How I can get your
answer automatically is more important than the answer.

The posted challenge would be more useful to symbolic
software (note that I have a vested interest in Axiom) if the
original challenge showed the failure of some symbolic
algorithms. The posted result would be more useful if it updated
an existing algorithm to handle the challenge. Such posts would
advance the state of the art.

So I don't have a problem with the idea of "challenges" but I do
feel that they are like "sunday crossword puzzle games", which
are entertaining but not adding to the state of the art.

Manuel Bronstein published a book on the algorithms for
integration. Given a challenge problem, show where his algorithm
fails and publish a change so that it can handle the challenge.

Where are the algorithms?

Tim Daly
Axiom Lead Developer

(your post is related to R J Fateman's posting, yes? My reader
does not show that ...)

I would bet a good part is taken using from Prudnikov et al.

On Jun 6, 12:00 pm, Axel Vogt <&nore...@axelvogt.de> wrote:

I would bet a good part is taken using
from Prudnikov et al.

"a good part"... how much is this?