Re: rationalize returns inexact results.

Re: rationalize returns inexact results.

[Anye Li]

Even if it is not a bug that rationalize returns inexact results 
(surprising and useless as they may be to me), I believe it is still a 
bug that (help rationalize) says "Return an exact number..." when in 
fact it does not.

Re: rationalize returns inexact results.

[Neil Jerram]

2008/7/26 Anye Li <li.anye.0@gmail.com>:
> Even if it is not a bug that rationalize returns inexact results (surprising
> and useless as they may be to me), I believe it is still a bug that (help
> rationalize) says "Return an exact number..." when in fact it does not.

Agreed.  I'm just fixing this now, to make (help rationalize) say the
same as the Guile manual says:

Returns the @emph{simplest} rational number differing
from @var{x} by no more than @var{eps}.

As required by @acronym{R5RS}, @code{rationalize} only returns an
exact result when both its arguments are exact.  Thus, you might need
to use @code{inexact->exact} on the arguments.

@lisp
(rationalize (inexact->exact 1.2) 1/100)
@result{} 6/5


Regards,
       Neil

Re: rationalize returns inexact results.

[Bill Schottstaedt]

> Since Guile now supports exact rationals, shouldn't it return an exact result?

I actually agree with you (I implemented rationalize in Guile, and originally
it always returned a ratio -- why else call it "rationalize"?), but Marius
pointed out that R5RS says "With the exception of inexact->exact, the 
operations described in this section must generally return inexact results
when given any inexact arguments" in the section including rationalize.
I doubt that R6RS fixed this bug.  So, to get an exact result you need to
give exact arguments:

guile> (rationalize (inexact->exact .33) 1/10)
1/3

rationalize returns inexact results.

[Anye Li]

[-- Attachment #1: Type: text/plain, Size: 394 bytes --]

Hi,

I've noticed that rationalize returns inexact results in Guile 1.8.3 (my
apologies if this has already been fixed), e.g. (rationalize 0.333333 0.01)
returns 0.333333333333333, which is inexact.  This is contrary to the claim
made by (help rationalize).  Since Guile now supports exact rationals,
shouldn't it return an exact result, e.g. 1/3 for the example given above?

Thanks,

Anye Li

[-- Attachment #2: Type: text/html, Size: 456 bytes --]