Numerical evaluation etc.



Hmm, perhaps I should have been more specific.

On Sat, Mar 20, 2004 at 09:21:53PM -0500, Stavros Macrakis wrote:
> > (C2) EV(exp(1.0+sqrt(-1.0)),NUMER);
> > The documentation claims this should be expanded ...
> 
> Starting with the symbolic expression:
> 
>   expr: exp(1+%i)$
> 
> you can get the numeric value in rectangular form:
> 
>   rectform(expr),numer;
> 
> or in polar form:
> 
>   polarform(expr),numer;
> 
> I agree that the documentation needs to be improved.  For that matter,
> the system's behavior could be more intuitive.  The problem here is that
> Maxima does not treat %i as a "number".

Thank you; this solves half my question, and lets me do the
calculation I was trying to do.  What I really wanted to ask was:

Suppose I have an expression whose value is a single definite number
(let's say a complex number).  Is there some command I can issue so
that Maxima will tell me a floating-point approximation to that
number, regardless of what the expression is?

Will rectform and polarform followed by ev(  ,numer) do this?

Given such a command, can it be applied to an expression with some
indeterminates to evaluate as far as possible?  (for example, a
polynomial might have all its coefficients completely evaluated in
floating-point)

I realize that there may be functions Maxima knows something about but
cannot evaluate numerically (perhaps only for some arguments).  But
that is surely not the case here...

I expect my background with MAPLE is showing here; MAPLE's "evalf"
function does exactly what I am describing, with a "digits" argument.
MAPLE's various shortcomings have encouraged me to seek an
alternative.  I gather bfloats are the answer to "digits", but for 
the moment I'm just trying to use Maxima like a better pocket calculator. 

Thanks,
Andrew