Re: [Chicken-users] IEEE float arithmetic

[felix winkelmann]

On 6/19/06, John Cowan <address@hidden> wrote:
> I'd like to have Chicken inexact numbers play nicely as IEEE floats,
> since on all practical platforms they *are* IEEE floats.  This involves
> the following modifications.
>
> 1) Allow the reader to understand Infinity.0, -Infinity.0, and NaN.0 .
> These should be defined in the reader as numeric literals.  As a fallback,
> they could be defined in Scheme code:
>
>         (define Infinity.0 (* 1.0e307 1.0e307))
>         (define -Infinity.0 (* -1.0e307 1.0e307))
>         (define NaN.0 (- Infinity.0 -Infinity.0))
>
> This, however, makes them only available in code, not in a Scheme datum.

Chicken already supports +inf, -inf and +nan.
(+inf.0, ... are supported as well, but a bug I just fixed
in the number-parser prevented it from working).

> 2) Remove the divide-by-zero trap in inexact division, and allow the IEEE
> results to appear openly.  This will mean that (/ 1.0 0.0) will evaluate
> to Infinity.0, (/ -1.0 0.0) to -Infinity.0, and (/ 0.0 0.0) to NaN.0.

Uh oh - I don't think this is good... Does it really give much to propagate
nan's in thise case through numeric operations?


cheers,
felix