Extension of the integrator - special functions


By adding more patterns to integrallookups, are you slowing down all 
integral processing (bad), or just the
case where other programs cannot find an integral (good)?

Have you looked at the "PMINT" code (poor man's integrator)  by 
Bronstein, to see if it can be inserted into
Maxima in some manner?  It seems to work, at least some of the time, on 
more special functions,
but is not fast, requiring simultaneous solution of systems of 
non-linear equations.  (It was discussed here with
respect to whether the copyright on the (short) program in Maple 
matters. I think it does not.)

RJF