On Saturday, January 28, 2017 at 8:31:50 PM UTC+1, parisse wrote: > > I would say it's easier to check that the gcd of A(x) and B(x+t) is not > trivial for the value of t that are integer roots of the resultant (with > other parameters replaced by 0). >
I'm sure there are more opportunites for optimizing further. The impact would not be as big as with finding the constant coeff itself. Note that replacing other parameters by 0 does not always work, for example > for sum((-1)^k*comb(n,k)/comb(k+a,k),k)=a/(n+a), I had to put non-0 values > for the parameter. > Proving the identity does not seem a problem here: sage: F(n,k)=(-1)^k*binomial(n,k)/binomial(k+a,k)/a*(n+a) sage: c = F(n,k).WZ_certificate(n,k); c (a + k)*k/((a + n)*(k - n - 1)) sage: G(n,k) = c * F(n,k) sage: (F(n+1,k) - F(n,k) - G(n,k+1) + G(n,k)).simplify_full() 0 (using a Pynac prototype and the #22090 branch) > I'd be curious to compare with the randomize evaluation method. > Is there a way to get the "Gosper term" of an expression from Giac? I can only see bool gosper(const polynome & P,const polynome & Q,const polynome & R,polynome & Y,gen & deno,GIAC_CONTEXT); -- You received this message because you are subscribed to the Google Groups "sage-devel" group. To unsubscribe from this group and stop receiving emails from it, send an email to sage-devel+unsubscr...@googlegroups.com. To post to this group, send email to sage-devel@googlegroups.com. Visit this group at https://groups.google.com/group/sage-devel. For more options, visit https://groups.google.com/d/optout.
