On 2012-01-10, at 19:40, Alex Shinn wrote, responding to me:
(re exact results)
That's not my understanding of the definition of `should' from RFC 2119;
the text reads
SHOULD This word, or the adjective "RECOMMENDED", mean that there may exist valid reasons in particular circumstances to ignore a particular item, but the full implications must be understood and carefully weighed before choosing a different course.
This sounds like a design decision (and that's the way I've always interpreted
`should'), rather than a specification of how a procedure should behave. I
know I'm being picky, but the original paragraph sounded wrong to me when I
read it carefully this morning, and I only picked on `should' after the fact.
It feels contradictory, even though I admit it isn't. I don't mind if my
suggestion isn't adopted, I just think the paragraph should be reworded.
(re inexact->exact and exact->inexact)
That's fine, but it's still a fact that the names of these procedures
are misleading. There's a distinction between a name like cdr, which is
meaningless to those who don't know and love the IBM 709 computer, and
inexact->exact, which might actively confuse the reader. I agree that
the procedures are correctly described in the entry, but the names are
at such variance from the description that the reader is justified in
wondering if it's the description that's wrong rather than the names.
Some text such as `Note: These procedures accept as an argument any
kind of number, despite their names.' would help clarify that. That
sounds kind of clunky (and somewhat duplicates the description), so
my preferred wording is `Note: The names of these procedures are
historical if not completely descriptive.', I don't think any
further explanation need be provided.
_______________________________________________ Scheme-reports mailing list Scheme-reports@x http://lists.scheme-reports.org/cgi-bin/mailman/listinfo/scheme-reports