On Sat, Feb 7, 2015 at 7:40 PM, John Cowan <xxxxxx@mercury.ccil.org> wrote:

> Arthur A. Gleckler scripsit:
>
> > This doesn't seem to be useful enough to warrant inclusion in a general
> > library.  Do you have some general class of applications in mind?
>
> It's the rational-number analogue of iota.  It's not as intuitive an
> ordering as 1, 2, 3, 4, ... and can't be used for indexing, but it
> does traverse the whole (non-negative) set of rational numbers.
> The particulars of the ordering are nothing special, except that
> they require only a single item of state, the last rational number
> generated, and there is a once-and-only-once guarantee.
>

Yes, but I'm still wondering what it would be used for.  Is it commonly
needed?