Fraction hackers...
Dwight Hughes
dwighth at ipa.net
Wed Sep 15 05:45:14 UTC 1999
The standard Squeak Floats are IEEE-754 double-precision (64 bit).
FloatArrays use IEEE 32 bit format. See the class comment in Float for
more info. Notice that there are two basic ways to create Fractions from
Floats: Float>>asApproximateFraction and Float>>asTrueFraction.
Float>>asFraction uses #asApproximateFraction - and that is what you
want if you want 0.2 asFraction = (1/5) to be true.
-- Dwight
DisneyLogic at disneyblast.com wrote:
>
> >[Florin Mateoc] Well, guess what:
> >
> > '(1/2)' + 0.5 breaks
> > (1/2) + '0.5' 1.0
> >
> > '(1/2)' * 0.5 breaks
> > (1/2) * '0.5' 0.25
> >
> > and so on...
> >
> > Plus , between
> > '(1/2)' = 0.5 "false"
> > (1/2) = '0.5' "true"
> > IMHO the second result is "truer" than the first one, so maybe we
> >should fix the first one instead of falsifying the second one if we care for
> >simmetry (I agree, simmetry is good)
>
> Is there some source or standard upon
> which the Squeak flonum arithmetic is based
> upon, like IEEE? Trying to define the
> "same as" or even the "near to" predicate
> for flonums is a tricky business.
>
> Thx, --jtg
>
> [snip]
More information about the Squeak-dev
mailing list
|