From: fateman@gmail.com   
      
   On Saturday, June 23, 2018 at 7:37:12 AM UTC-7, j4n bur53 wrote:   
   > Funny new system on the scene "yacas",   
   > yet another computer algebra system.   
   >   
   Yacas has been around since at least 2002. Does this   
   count as "new"?   
      
   The particular limit problem is probably not the most   
   valuable testing object.  What does Sum(f(n), n=1..k)   
   mean for non-integer k?  There are various definitions   
   for limit, but the epsilon-delta definition, which is   
   implicit in most of the integration programs in CAS,   
   works for continuous functions, which this sum is not.   
   Some sums can be analytically continued, because they are,   
   for example, geometric series as is the numerator of this   
   example.   
      
      
   While running out of stack is not a great way to respond   
   to a question, especially if there is a proper answer,   
   Yacas is not alone in its fruitlessly suggesting that   
   maybe the answer is computable, but so far as it has   
   been able to tell, you might need more stack to find it.   
      
   Mathematica used to make such a response to, I recall,  z=z+1;   
   not so, anymore. So   
   I just typed f[i_]:=f[i-1];  f[0]  and got Hold[f[-4095-1]]   
   which most people might consider peculiar. Is it a bug?   
   Should Mathematica say something about stack overflow,   
   the way it used to?  What if you also said, f[-4094]=gotcha; f[0]  f[2]   
   what do you suppose the two answers should be?   
      
   Anyway, making a "whataboutery" argument about criticisms of   
   Yacas is not the point.  Rather it is the example.   
   Spock [ibid]   
      
   --- SoupGate-Win32 v1.05   
    * Origin: you cannot sedate... all the things you hate (1:229/2)