changing floating point number input.. (was realroots...)
Subject: changing floating point number input.. (was realroots...)
From: Stavros Macrakis
Date: Tue, 29 Aug 2006 11:18:56 -0400
On 8/29/06, Robert Dodier <robert.dodier at gmail.com> wrote:
> (%o7) float_inf(1, 3CC, 10000000000000)
> (%o8) float_inf(- 1, 3CC, 10000000000000)
> (%o9) float_nan(- 1, 3CC, 18000000000000)
>
> Printing the sign, exponent, and significand makes the
> result theoretically readable ....
Nice. But aren't most of the digits reported there redundant? If
not, what useful information do they give to the user (well, at least
the user who isn't an IEEE float geek)? I would rather have these
print as float_inf(32), -float_inf(32), and float_nan(xx), where 32 is
the precision of the float, and xx is some something telling you what
sort of NaN it is, preferably a string or identifier, not a number.
-s