On Tue, 2009-09-08 at 03:33 -1000, Shiro Kawai wrote:
> From: Adrien_"Pied"_Piérard <[email protected]>
> > And to get things done, I need hashtables.
> Perhaps that's one possible way a standard can be. But if what
> you want are those things, you don't need to wait for the standard.
> You can just pick an implementation today that suits your
> purpose.
First of all, to Mssr. Piérard; hash tables already exist
in a standard form. You should look for schemes which
implement SRFI-69 and, if your needs are esoteric, SRFI-90.
Which is, I think, nearly all of them because those libraries
are written in portable scheme. I'm fairly sure that
SLIB provides hash tables too, and many implementations
can use SLIB.
But Mssr. Piérard has demonstrated something; existence
as SRFI's and third-party libraries is not sufficient.
Particularly for a language like scheme with a small
community, people look to the standard document, and if
they don't find it there they assume that it does not
exist. And nonexistence of basic utilities quickly
develops into a vote of no confidence in the language
and its community, no matter how easy it is to implement
those utilities in that language.
Hash tables really are ripe for a standard. They're darned
useful to programmers, already widely implemented, and
already have a standard interface between implementations.
They're a classic example of "consensus exists already, and
it's the job of the standard to document the consensus."
I think that the next report ought to include a description
of hash tables as a standard library. For "large" scheme
definitely; probably for "small" scheme as well.
Bear
_______________________________________________
r6rs-discuss mailing list
[email protected]
http://lists.r6rs.org/cgi-bin/mailman/listinfo/r6rs-discuss
