changing floating point number input.. (was realroots...)
Subject: changing floating point number input.. (was realroots...)
From: Robert Dodier
Date: Sun, 27 Aug 2006 16:37:04 -0600
On 8/26/06, Richard Fateman <fateman at cs.berkeley.edu> wrote:
> Effects: if you put too many digits in a number like
> 1.234567890123456789012345e0, then you will get a bigfloat.
I think that auto-conversion would get annoying sooner or later.
I'm not sure I would want to automatically convert 1e600 to a
bigfloat, either.
I think I would be more interested if bigfloat was reworked
to make it an extension of IEEE 754, i.e. for x within the
range of ordinary floats, float(x) is identical to bigfloat(x).
As it stands, bigfloats are different enough to make automatic
conversions a potential source of unintended consequences.
FWIW
Robert