Translated code and simplification of %pi in presence of numer=true
Subject: Translated code and simplification of %pi in presence of numer=true
From: Stefano Ferri
Date: Mon, 20 Apr 2009 10:30:52 +0200
Hi,
just to know... could this problem also related to sf bug 2477795 "assume:
problems with fractions or multiples of %pi and %e"?
Some time ago Dieter Kaiser had a look at the involved code,ad sent this
message:
http://www.math.utexas.edu/pipermail/maxima/2009/015344.html
It's not the same thing, but also here there was the proposal of numerical
evaluation irrational constants to solve a piece of that problem...
Stefano
In data luned? 20 aprile 2009 02:56:18, Richard Fateman ha scritto:
: > The conversion happens inside timesin, which explicitly checks for $%e.
>
> What a mess.
>
> Philosophically speaking, the "simplifier" should not be changing 2*pi
> to 6.28...
> since those two items are not equal. One is an approximation of the other.
>
> A better design, suggestion anyway:
>
> There should be another program, probably NOT intertwined with the
> simplifier, that does floating-point evaluation.
>
> It might be meval, {with numer=true} but that's a rat's nest too.
>
> It could be float_eval, or maybe just $float and could be modeled on
> the $bfloat program.
>
> RJF
>
> Barton Willis wrote:
> > -----maxima-bounces at math.utexas.edu wrote: -----
> >
> >
> >
> >
> > The %e --> 2.718 conversion seems to happen inside simptimes, not
>
> _______________________________________________
> Maxima mailing list
> Maxima at math.utexas.edu
> http://www.math.utexas.edu/mailman/listinfo/maxima