Oh, I see it now. gradef() and diff() got it right, I was wrong, it?s the chain rule.
Rich
From: Barton Willis
Sent: Thursday, July 11, 2013 6:23 PM
To: Richard Hennessy
Cc: maxima at math.utexas.edu
Subject: RE: [Maxima] gradef problem, was pw.mac problem
I don't see a problem with diff(pwdelta(s*x-v),x,4) --> diff_pwdelta(4,s*x-v)*s^4 . This seems to be a correct application
of the chain rule.
--Barton
--------------------------------------------------------------------------------
We have,
gradef(pwdelta(x), diff_pwdelta(1,x))$
gradef(diff_pwdelta(n,x), 'und, diff_pwdelta(n+1,x))$
<snip>
diff(pwdelta(s*x-v),x,4); // problem below
-> diff_pwdelta(4,s*x-v)*s^4
This last line should be
diff(pwdelta(s*x - v), x, 4);
-> diff_pwdelta(4, s*x-v)
by definition of diff_pwdelta().
I don?t know a workaround for this one.
Rich