Re: numerical conditioning MAGNITUDE and / William D Clinger (21 Jun 2006 13:29 UTC)
Re: numerical conditioning MAGNITUDE and / John Cowan (21 Jun 2006 13:38 UTC)
Re: numerical conditioning MAGNITUDE and / Aubrey Jaffer (21 Jun 2006 17:32 UTC)

Re: numerical conditioning MAGNITUDE and / William D Clinger 21 Jun 2006 13:29 UTC

Aubrey Jaffer wrote:
> Writing arithmetic routines in Scheme does not eliminate the need to
> do the proper numerical analysis....A correct implementation is:

Thank you, Aubrey.

Aubrey Jaffer has kindly given permission for his code to go
into the reference implementation.

> Does SRFI-77 (or R5RS) mandate that floating-point procedures work
> for arguments generating the full range of possible outputs?

No.  Neither mandates that inexact reals be represented as
binary floating point, or that inexact reals follow the IEEE
standards for binary floating point, or that implementations
of the standard procedures be of high quality.

Will