Translated code and simplification of %pi in presence of numer=true



In data luned? 20 aprile 2009 16:44:58, Richard Fateman ha scritto:
: > If you use numerical approximations in assume,  then you are inserting
> potentially false assertions into Maxima.
> e.g.
> assume(x>=%pi),numer
>
>
> is  (x   >= %pi + 10^(-20))   might return true even if it is not
> necessarily true mathematically.
>
> If you want to have a database of assertions which are sometimes true,
> but not always, maybe numerical
> evaluation is OK.  Maybe if the assertions already have floating point
> numbers it might be OK.
> However, %pi is not a floating point number.
>
> RJF

You're right, I didn't think about that...
Thanks
Stefano