patterns for integration



Richard, thank you for the answer.

Richard Fateman wrote:
> 
> If you want the program to just flounder around for a while trying all
> possible ways of matching something, don't expect it to be efficient or even
> effective.

Yes, I know that this is not efficient. But Im trying to do at least 
something. And I hope that I find also some more efficient way.

Would be learning from integration routines of Maxima useful to get 
better method for my purposes? Or must I study topics on symbolic 
integration to solve this problem?

What I wish is to write function or something, which gives hints 
concerning suitable method for integration, like the function hint in 
Maple. We should cover at least basic formulas and typical integration 
by parts (seems to be easy - but, as you said, expensive for 
calculations, the first attempt is already on 
http://old.mendelu.cz/~marik/maw/index.php?lang=en&form=integral ), 
further we plan to cover basic substitutions x=a*tan(t), x=a*sin(t), 
x=a*cos(t), cos(x)=t, sin(x)=t, cos(x)=t, tan(x)=t, ax+b=t^2 and 
Ostrogradskii method for polynomial/sqrt(quadratic polynomial). And this 
should cover basic course in calculus on technical university.

I tried to put trace command to changevar and evaluate integral in 
Maxima to see, if changevar is used when integrating in Maxima but (as I 
supposed) without any success.

Robert Marik

> You seem to want some kind of unification program, rather than matching, and
> that is extremely expensive to run.
> 
> RJf
>  
>