On Sep 2, 2009, at 7:40 PM, Raymond Toy wrote:
>
> Does anyone actually compile maxima with flonum other than the default
> double-float? I don't.
>
> Ray
>
Me neither. Probably nobody outside developer community knows about
this. I didn't until recently. Also, this is possible only on CMUCL
and CLISP. Considering all these factors I think it's highly probable
that nobody uses it...
Perhaps strengthening bigfloats (I have no idea how except by using a
C library..) is more sensible than supporting various 'flonums'. I
think double-floats (which are fast) have more than enough precision
for all approximate tasks..
Regards,
Ziga