I know of no user-level way to tell Maxima that diff(c[k],c[l]) = kron_delta(k,l).
(The Kronecker delta function.) As far as I know, pattern matching and tellsimp rules
cannot do this. Even if you were able to tell Maxima that diff(c[k],c[l]) = kron_delta(k,l), you
would need to help Maxima simplify summands that involve the Kronecker delta:
(%i16) e : sum(a(k) * kron_delta(k,l),k,minf,inf)$
(%i17) e,simpsum;
(%o17) (sum(a(k)*kron_delta(k,l),k,0,inf))+sum(a(-k)*kron_delta(-k,l),k,1,inf)
(%i18) simplify_sum(e);
(%o18) (sum(a(k)*kron_delta(k,l),k,0,inf))+sum(a(-k)*kron_delta(-k,l),k,1,inf)
--Barton
-----maxima-bounces at math.utexas.edu wrote: -----
>Say?i?have?f(x):=sum(c[i]*x,i,0,N);
>I'd?like?to?diff(f(x),c[0]). But?it?always?returns?0.
>How?should?I?do?that?