[Date Prev][Date Next][Thread Prev][Thread Next][Date Index][Thread Index]

Re: reading NaNs

This page is part of the web mail archives of SRFI 77 from before July 7th, 2015. The new archives for SRFI 77 are here. Eventually, the entire history will be moved there, including any new messages.



   Date: Wed, 26 Oct 2005 19:02:44 -0500
   From: Alan Watson <a.watson@xxxxxxxxxxxxxxxx>

   Marcin 'Qrczak' Kowalczyk wrote:
   > There is no other model competing for being the default. This means
   > that it's requesting other models which should be explicit.

   I agree.

   Selecting different models becomes quite easy if you push inexacts off 
   into the library and have a half-decent module system.

Why is this not equally applicable to IEEE flonum arithmetic?  If what
you *want* is IEEE flonum arithmetic, then by all means say so in your
program; but if you don't care, why should IEEE arithmetic be mandated
as the default, precluding, for example, arbitrary-precision flonum
arithmetic, with some user-frobbable parameter specifying the actual
precision, as I mentioned CLISP & SISC support? or, say, in systems
oriented toward business accounting applications, a ten-based flonum
arithmetic system?