From: fateman@cs.berkeley.edu   
      
   I wrote, in private email to Jan....   
      
   Just think, you are reporting a bug that only you could plausibly fix.   
     This is a bug in a program that you have written.   
   I suggest you set up a newsgroup for people who use your program and   
   report bugs there.   
   I don't expect you to take my advice, but maybe you could   
   reduce your posting activity to sci.math.symbolic.   
      
   ...........   
   I figured that Jan is  off his meds, but nevertheless hoping to give him   
   advice. Oh well.   
      
   I would not say that special GCD algorithms are universally better   
   than GB.  I suppose someone could write a really bad polynomial   
   GCD algorithm that would be exceptionally slow.  Fair benchmarks   
   are difficult to construct.  In particular, for a CAS, it is often possible   
   for an experienced user of a system to identify a better command   
   or sequence of commands to obtain a result than might be chosen   
   by a naive user.  (e.g. in Maxima, using rat() rather than expand()).   
      
   Nevertheless, there are good reasons to believe that,   
   in any reasonably implemented CAS which   
   has  polynomial GCD and GB procedures,   that using GB to   
   implement a GCD  would be faster than the GCD.   
      
   --- SoupGate-Win32 v1.05   
    * Origin: you cannot sedate... all the things you hate (1:229/2)