Sudarshan S Chawathe scripsit:

> This answers that question.  It is probably a reasonable design choice,
> and certainly the prerogative of the SRFI author.  However, it would be
> very helpful if the SRFI were very explicit in this regard.

I have strengthened the wording in the definition section
and other places in the draft.

> I will mention also that it may be useful to consider comparators that
> reflect underlying partial (not total) orders.  A textbook use case
> would be using a sorting library (that uses comparators) to sort plane
> tickets that are only partially ordered by preference: x is better than
> y if x is both a cheaper and a shorter flight than y.

Fair enough, but I think it's an unreasonable burden on implementers to
force them to make the procedures specified in the SRFI to work
correctly if there is only a partial order.

--
John Cowan          http://www.ccil.org/~cowan        xxxxxx@ccil.org
The Imperials are decadent, 300 pound free-range chickens (except they have
teeth, arms instead of wings, and dinosaurlike tails).  --Elyse Grasso