From: fateman@cs.berkeley.edu   
      
   On 7/12/2017 7:49 AM, bursejan@gmail.com wrote:   
   > You mean domain univariate polynomials possibly rational   
   > function coefficients right? Which is not what I want to do.   
      
   No, I do not mean polynomials over  Q(x2,x3, ...)[x1].   
   if that is the intention of your question.  The usual most   
   useful GCD is over Z[x1, ...,xn].   
   >   
   > And not multivariate polynomials with value coefficients?   
     .   
   > Which is what I want to do, and which is at stake.   
   What is a "value coefficient" ?   do you mean Z[x1, ...,xn]?   
   >   
   > Can you show us here and now such an algorith, that   
   > gives something else than 1 for gcd(x*y^2-y,y*x^2-x)?   
   >   
   in Maxima, typing gcd(x*y^2-y,y*x^2-x);   returns   x*y-1   
      
   It doesn't show the internal algorithm, of which there are   
   several.  If you want to see the algorithm you will have   
   to look at the source code.  The line above only   
   shows you the command and the result.   
      
      
      
   > Hint: You need only to paste your link to   
   > http://maxima-online.org/ here in this newsgroup.   
      
   Nice.  I have never tried this.   
      
   http://maxima-online.org/#?in=%20gcd(x*y%5E2-y%2Cy*x%5E2-x)%3B%0A%20%09%0A   
      
   will show you exactly that answer.   
      
   to see gcd options, look at   
      
   http://maxima-online.org/#?in=%20%3F%20gcd%3B%0A%0A%20%09%0A   
      
   >   
   >> I DON"T NEED ANY REPLACEMENT. There are several   
   >> algorithms for multivariate GCD which incorporate   
   >> various clever ideas developed since the mid 1960s.   
      
   --- SoupGate-Win32 v1.05   
    * Origin: you cannot sedate... all the things you hate (1:229/2)