RE: PLT's intent to support SRFI's?

[Matthew Flatt <mflatt@cs.utah.edu>]

Quoting "Will Fitzgerald":
> Could you introduce a standard process so that SRFIs could be easily
> included by others who might have the time? Perhaps standardizing on a srfi
> folder in the collects directory, or a REQUIRE-SRFI procedure?

How about a "srfi" collection where each SRFI has it's own numerically
named sub-collection? Each SRFI would be implemented by a `module',
which can provide both syntax and procedures.

To get SRFI 5:

  (require (lib "srfi.ss" "srfi" "5"))

Since that's somewhat verbose, MzLib could include a "srfi.ss" library
that provides `require-srfi':

   (require (lib "srfi.ss"))
   (require-srfi 5)


[Shriram K. and Mike S.: Did I forget anything from our earlier
discussion about SRFI collections?]

> Perhaps
> setting up a template for units/signatures so that PLT SRFIs get built in
> the same way?

The new `module' form makes this simpler. Each "srfi.ss" file would
have the form

  (module srfi <lang> ; where <lang> is probably `mzscheme'
    <expression or definition>
    ...
    (provide <id> ...))

And we'll provide a tool that makes .plt file creation easier (perhaps
as a new blade on the Swiss-army mzc).

Matthew