[plt-scheme] Are new Schemers supposed to be reading SRFIs?

From: Geoffrey S. Knauth (geoff at knauth.org)
Date: Mon May 14 07:02:58 EDT 2007

On May 10, 2007, at 15:27, John Clements wrote:
> In some ways, this resembles the problems that doctors have; in  
> order to give the best possible advice, they need to keep up with  
> all of the most recent findings, while at the same time discounting  
> somewhat the more recent & unverified one.
> I don't think that CS has yet evolved (or, hitherto, needed to  
> evolve) a good channel for this kind of dissemination.

If there were contracts or machine readable documentation everywhere,  
we could ask the computer, "Has anyone written code that takes types  
X Y and produces Z?"  A smart network would find implementations in  
different languages, rate them, and offer to help convert them into  
our favorite language.

For DrScheme there's Help Desk and browsing the collects  
directories.  Maybe Help Desk could add type matching to its skills.   
The Computer Language Benchmarks Game (language shootout) is fun to  
read, and Google Code will some day give better answers.

On May 11, 2007, at 07:09, Jens Axel Søgaard compiled a cheat-list:
>   SRFI 1   (lists)
>   SRFI 13  (strings)
>   SRFI 19  (time and date)
> Also in use:
>   SRFI 14  (characters)
>   SRFI 27  (random numbers)
>   SRFI 42  (eager comprehensions)
>   SRFI 43  (vectors)
> Seen, but not often:
>   SRFI 26  (cut - specialiazing)
>   SRFI 67  (compare procedures)

I thought, "Good advice from a knowledgeable person!"  I wondered how  
long Jens Axel took to assemble that list, and whether an automated  
mechanism in DrScheme should report what and how often things get  
used.  PLaneT could report how popular packages are.  I wonder what  
library science thinks of computer science.


Posted on the users mailing list.