Ralf Hemmecke wrote:
>
> But I guess that is a follow-up error from the computation of
> f3:=D(f2,x). Also there I wouldn't want to see %B to appear in the output.
>
> I faintly remember that Waldek introduced that local variables to
> indicate "derivation wrt the n-th variable", because there was a bug for
> another expression. Looks to me like this code must be revisited.
Actually, use of dummy variables comes from NAG version. NAG folks
used fixed names of dummy variables, but this may cause trouble.
So I changed code to use fresh dummies. At the same time I modified
code to make it more efficient. As you saw I forgot to include
one of conditions for equality in the new code. Once equality
of kernels is wrong one can get all kinds of weird results:
given kernel may be replaced by different unequal kernel.
For derivatives natural representation would be dervative
operator applied to kernel. Unlike some other systems AXIOM/FriCAS
uses positional notation for derivatives, that is we need to know
how many times we differentiate with respect to each argument.
At low level arguments to '%diff' could be original kernel and
list of integers giving order of derivatives. Unfortunately,
due to type rules arguments to kernels must be Expression-s and
in finite characteristic natural homomorphizm from integers to
expressions is not injective. So, instead we use current funky
notation. If we could attach list of integers to a kernel, then
we could use much simpler representation.
--
Waldek Hebisch
[email protected]
--
You received this message because you are subscribed to the Google Groups
"FriCAS - computer algebra system" group.
To unsubscribe from this group and stop receiving emails from it, send an email
to [email protected].
To post to this group, send email to [email protected].
Visit this group at http://groups.google.com/group/fricas-devel.
For more options, visit https://groups.google.com/groups/opt_out.
