On Mon, Jul 27, 2009 at 4:05 PM, Dieter Kaiser <drdieterkaiser at web.de>wrote:
>
> ...But the result is not very well simplified. It contains sin^2+cos^2
> terms and I think the atan2 terms should vanish too. More work can be
> done.
>
Fortunately, trigsimp gets sin^2+cos^2 => 1
It would be nice if we could improve atan2 simplification in general, but
there's a problem:
Usually when you see terms like atan2(sin(x),cos(x)), the substantively
correct result is x, that is, the analytic continuation of the -%pi < x <
%pi case.
The problem with this is of course that atan2 is defined to have a range of
-pi .. pi ....
-s