From: Albert_Rich@msn.com   
      
   On Tuesday, January 13, 2015 at 11:47:51 PM UTC-10, Nasser M. Abbasi wrote:   
      
   > If you like to share the antiderivative scoring function   
   > you used, I will use it in my program. I will call it,   
   > with the API it uses, and use whatever scoring   
   > number it returns. If this function is already at   
   > your web site, I must have missed it.   
      
   Actually I graded the Timofeev test suite results manually by inspection   
   simply looking for the imaginary unit, etc.  However, since you wish to grade   
   1000s of problems obviously the grading should be automated.  This can be   
   accomplished using    
   Mathematica's built-in FreeQ function (presumably Maple has something   
   similar).  For example, if the Boolean   
      
      Not[FreeQ[result,Complex]] && FreeQ[optimal,Complex]   
      
   returns True, the result unnecessarily contains the imaginary unit, and hence   
   is non-optimal despite how small in size it might be.   
      
   BTW, my other idea to show the average ratio of result to optimal sizes would   
   mean the executive summary could make more meaningful statements like    
      
   "Antiderivatives returned by system X are on average 2.5 times the optimal   
   size."   
      
   instead of just giving raw total leaf counts.   
      
   FWIW,   
   Albert   
      
   --- SoupGate-Win32 v1.05   
    * Origin: you cannot sedate... all the things you hate (1:229/2)