[Scheme-reports] exact-closed in features list Per Bothner (14 Oct 2014 21:22 UTC)
Re: [Scheme-reports] exact-closed in features list John Cowan (14 Oct 2014 21:46 UTC)

Re: [Scheme-reports] exact-closed in features list John Cowan 14 Oct 2014 21:41 UTC

Per Bothner scripsit:

> Frankly, this seems weird.  I would expect / to give exact values
> given exact inputs if an implementation has this property.

Implementing / requires support for rationals, whereas implementing
support for +, -, * requires only support for bignums.  (Of course,
technically no operation is closed over even the integers because of
implementation restrictions.)

> Wikipedia defines "Algebraic operation" to include "taking roots".

That was an oversight on my part.  I'll go with a variant of your
second wording:

>    The algebraic operations +, -, *, and expt where the second
>    argument is a non-negative integer all produce exact values given
>    exact inputs.

--
John Cowan          http://www.ccil.org/~cowan        cowan@ccil.org
Knowledge studies others / Wisdom is self-known;
Muscle masters brothers / Self-mastery is bone;
Content need never borrow / Ambition wanders blind;
Vitality cleaves to the marrow / Leaving death behind.    --Tao 33 (Bynner)

_______________________________________________
Scheme-reports mailing list
Scheme-reports@scheme-reports.org
http://lists.scheme-reports.org/cgi-bin/mailman/listinfo/scheme-reports