Le 20/04/2012 19:55, Stavros Macrakis a ?crit :
> Alas, Maxima is not smart enough to realize that acos(sin(x)) is real
> for real x. In fact, Maxima doesn't even know that -1 <= sin(x) <= 1
> for real x. (By the way, Maxima treats undeclared variables as real in
> most cases....)
>
> I'd think the right way to handle that would be (symbolic) interval
> analysis where sin(x) => [-1,1] for real x and acos([-1,1]) => [0,%pi]
> (assuming the principal value interpretation).
Yes.
By the way, does Maxima handle variable domain ?
I mean a way to declare (or assume) that variable x is always in set s.
> But Maxima doesn't
> currently do that, sorry.
>
OK. Thanks anyway.
Jean
> [...]