There is a flag in commercial macsyma, newdiff,
which can be set to true in which case
diff(f(2*x),x) comes out as 2*f^(1)(2*x) where ^(1) means
diff wrt first argument.
RJF
C
Nikolaos I. Ioakimidis wrote:
> Dear colleagues,
>
> I try to verify the accuracy of the solution
>
> y(x,t) = f(x+c*t)+g(x-c*t)
>
> to the classical wave partial differential equation. (Analogously,
> with the Laplace equation.) I prepared the code and
> found the solutions (such as the above one for the wave
> equation), but now I wish to proceed with the verification
> through differentations by using Maxima.
>
> My impression would be that Maxima can find that
>
> (d/dx)f(2*x) = 2*f '(2*x),
>
> but it failed. Therefore the above verification and similar
> verifications cannot succeed since Maxima seems not
> knowing the chain rule for not concrete functions (contrary
> to Maple and Mathematica). Is this true with my command
>
> diff(f(2*x),x)
>
> that I tried to use or some trivial error (due to me) is
> present in this so simple Maxima command?
>
> It is understood that the differentiation algorithms of
> Maxima can be easily extended if required, but, never-
> theless, I would prefer that the chain rule in the present
> case could also be available to Maxima in advance.
>
> Your help will be greatly appreciated.
>
> Many thanks in advance and best regards from Patras,
>
> Nikos
>
>
>
>
>
>
> _______________________________________________
> Maxima mailing list
> Maxima@www.math.utexas.edu
> http://www.math.utexas.edu/mailman/listinfo/maxima
>