On Thu, Apr 12, 2007 at 04:29:51PM -0700, Richard Fateman wrote:
> If you think that by hiding some number of digits in output you will
> avoid all differences between decimal and binary,
> I think you will be disappointed. The examples that illustrate the
> "strange" behavior will be less common, but they will exist and be, in
> some sense, even harder to explain.
I acknowledge and agree that binary floats and decimal floats will
never act the same simply by reducing the printed precision. However I
think users unfamiliar with floating point will be less often confused
by the results printed when fpprintprec = 6 (or 7 or 8). It still
won't prevent problems like
is(1.2^2 * 100 - 144 = 0);
false
But less often confused is helpful and fpprintprec:6 even seems to fit
with serious users needs as well.
--
Daniel Lakeland
dlakelan at street-artists.org
http://www.street-artists.org/~dlakelan