I suggest that atan(tan(x)) not simplify unless we know -%pi/2 <x <%pi/2
This is what Macsyma appears to do.
RJF
----- Original Message -----
From: "Robert Dodier" <robert.dodier at gmail.com>
To: "Maxima" <maxima at math.utexas.edu>
Sent: Monday, March 13, 2006 7:52 AM
Subject: simplification of atan(tan(x)), etc
> hello,
>
> maxima simplifies atan(tan(x)) to x for any non-float x.
> this is reported as SF bug # 1448605 .
> (http://sourceforge.net/tracker/index.php?func=detail&aid=1448605&group_id=4933&atid=104933)
>
> shouldn't x be pulled back into the range [- %pi/2, %pi/2] ?
>
> SIMP-%ATAN is in src/trigi.lisp. SIMP-%ASIN and other trig inverses
> are in src/trigo.lisp. SIMP-%ASIN etc all implement the same kind of
> simplification (afoo(foo(x)) => x).
>
> comments?
>
> robert dodier
>
> _______________________________________________
> Maxima mailing list
> Maxima at math.utexas.edu
> http://www.math.utexas.edu/mailman/listinfo/maxima
>