From: fateman@cs.berkeley.edu   
      
   On 3/4/2014 9:51 AM, hebisch@math.uni.wroc.pl wrote:   
   >   Users may want to use more forms,   
   > and integration program should express (if possible) answer   
   > in terms of functions appearing in users input.   
      
   I agree with this.   
      
   There is a larger question, which is the context of an actual   
   integral: to what extent does it matter if one has a   
   closed form or not?  It is pressing if the integral   
   is actually a multiple (nested) integral as appeared   
   in evaluating Feynman diagrams, in which case evaluating   
   some of the inner ones in closed form reduced the   
   computation substantially.  The Reduce system was used   
   (heavily?) for this. But for a single integral, the case is not so clear.   
      
   Today, I think that most computational scientists will first turn to a   
   numerical quadrature library for a (definite) integral. The insight that   
   comes from viewing a graph with respect to some parameter may not  be a   
   total substitute for the insight in using a formula, but when the   
   formula is full of unfamiliar functions, it is hard to understand.   
    From this perspective, an approximation by series (taylor, asymptotic)   
   might be preferable.   
      
   Also, it may be the case that (after considerable computation) the   
   computer algebra system can't figure out some formula, and the   
   computation must be done numerically anyway.   
      
   It is also possible that the formula is difficult to evaluate   
   numerically (if that is its ultimate use)... large, unstable...   
      
   So the integration "problem" as envisioned by Liouville may   
   be more pure mathematics than applied mathematics. That is,   
   as the subject line suggests,  integral for fun.   
   RJF   
      
   --- SoupGate-Win32 v1.05   
    * Origin: you cannot sedate... all the things you hate (1:229/2)