inexact->exact in 1.7.1

inexact->exact in 1.7.1

[Rick Taube]

this behavior of inexact->exact in guile 1.7.1 causes my program to 
become confused:

guile> (inexact->exact 48.0)
48
guile> (inexact->exact (* .1 480.0))
48
guile> (inexact->exact (* (- 1.1 1) 480.0))
3377699720527875/70368744177664

guile>(quit)
$ uname -a
Darwin pinhead.music.uiuc.edu 7.8.0 Darwin Kernel Version 7.8.0: Wed 
Dec 22 14:26:17 PST 2004; root:xnu/xnu-517.11.1.obj~1/RELEASE_PPC  
Power Macintosh powerpc


-rick


i attempted to see if guile 1.7.2 fixes this but it doesnt build on 
darwin. perhaps someone could tell me



_______________________________________________
Guile-user mailing list
Guile-user@gnu.org
http://lists.gnu.org/mailman/listinfo/guile-user

Re: inexact->exact in 1.7.1

[Per Bothner]

Rick Taube wrote:
> this behavior of inexact->exact in guile 1.7.1 causes my program to 
> become confused:

I suspect you're the one who is confused.
You need to learn about floating-point arithmetic.
(- 1.1 1) does not necessarily give the same result as 0.1.

> i attempted to see if guile 1.7.2 fixes this

Why do you think there is a bug?
-- 
	--Per Bothner
per@bothner.com   http://per.bothner.com/


_______________________________________________
Guile-user mailing list
Guile-user@gnu.org
http://lists.gnu.org/mailman/listinfo/guile-user

Re: inexact->exact in 1.7.1

[Marius Vollmer]

Rick Taube <taube@uiuc.edu> writes:

> this behavior of inexact->exact in guile 1.7.1 causes my program to
> become confused:
>
> guile> (inexact->exact 48.0)
> 48
> guile> (inexact->exact (* .1 480.0))
> 48
> guile> (inexact->exact (* (- 1.1 1) 480.0))
> 3377699720527875/70368744177664

This happens because Guile now has exact rationals.  This NEWS entry
explains the behavior:

    ** inexact->exact no longer returns only integers.

    Without exact rationals, the closest exact number was always an
    integer, but now inexact->exact returns the fraction that is exactly
    equal to a floating point number.  For example:

        (inexact->exact 1.234)
        => 694680242521899/562949953421312

    When you want the old behavior, use 'round' explicitely:

        (inexact->exact (round 1.234))
        => 1


In your case, "1.1" is not exactly 11/10 since IEEE double can not
represent that number exactly.  You can write "#e1.1" instead, if you
want an exact 11/10:

    guile> #e1.1
    11/10
    guile> (inexact->exact (* (- #e1.1 1) #e480.0))
    48

A IEEE double can not represent 0.1 exactly either, and you just got
lucky that (* 0.1 480.0) is exactly 48:

guile> (- (* 0.1 480.0)
          (* (- 1.1 1) 480.0))
-4.2632564145606e-14


_______________________________________________
Guile-user mailing list
Guile-user@gnu.org
http://lists.gnu.org/mailman/listinfo/guile-user

Re: inexact->exact in 1.7.1

[Rick Taube]

Thank you, i was porting my code from 1.6.7 to the new version where it 
broke and I forgot about this. time for a m-x tags-search!
--rick


> This happens because Guile now has exact rationals.  This NEWS entry
> explains the behavior:
>
>     ** inexact->exact no longer returns only integers.
>
>     Without exact rationals, the closest exact number was always an
>     integer, but now inexact->exact returns the fraction that is 
> exactly
>     equal to a floating point number.  For example:
>
>         (inexact->exact 1.234)
>         => 694680242521899/562949953421312
>
>     When you want the old behavior, use 'round' explicitely:
>
>         (inexact->exact (round 1.234))
>         => 1
>
>
> In your case, "1.1" is not exactly 11/10 since IEEE double can not
> represent that number exactly.  You can write "#e1.1" instead, if you
> want an exact 11/10:
>
>     guile> #e1.1
>     11/10
>     guile> (inexact->exact (* (- #e1.1 1) #e480.0))
>     48
>
> A IEEE double can not represent 0.1 exactly either, and you just got
> lucky that (* 0.1 480.0) is exactly 48:
>
> guile> (- (* 0.1 480.0)
>           (* (- 1.1 1) 480.0))
> -4.2632564145606e-14



_______________________________________________
Guile-user mailing list
Guile-user@gnu.org
http://lists.gnu.org/mailman/listinfo/guile-user