[Date Prev][Date Next][Thread Prev][Thread Next][Date Index][Thread Index]

Re: [Scheme-reports] [scheme-reports-wg1] John Cowan votes for `r6rs` (operational equivalence)

On Mon, Nov 26, 2012 at 10:35 AM, John Cowan <cowan@x> wrote:
In the matter of the definition of `eqv?` on inexact numbers, I
have voted for `r6rs` (operational equivalence), `representational`
(representational, i.e., bitwise equivalence), and `r5rs` (numerical
equality as in `=`) in that order.  Note that `eqv?` will remain
unspecified on two NaNs no matter how this vote comes out.

I've voted this way:
Rationale: The representational definition clearly and concretely defines what an implementer should do to implement eqv? for inexact numbers, which makes it better than the r6rs option, and it takes into account issues like precision, which makes it better than the r5rs option. The r6rs option is better than the r5rs option for the same reason. Thus my ordering of preferences.
Scheme-reports mailing list