Email list hosting service & mailing list manager


Re: Small modification Richard Kelsey 13 Jul 1999 00:08 UTC

   Date: Mon, 12 Jul 1999 16:35:01 -0400
   From: Olin Shivers <xxxxxx@mongkok.ai.mit.edu>

   [...] It's better to make an effort to make each SRFI represent
   some consensus, and to hash out a design that is as much as possible
   The Right Thing.

I agree with you that it is better to avoid overlapping SRFIs.

But...

It is July, 1999.  Scheme does not have records.  I have given up
waiting for a consensus.  I thought that the whole point of the SRFIs
was to make it possible to move forward without one.

You want an extensible syntax.  Matthias wants initialized fields.
There is no end to what a record definition can be made to do.  Look
at Common Lisp's DEFSTRUCT.  I do not believe that we can come to
a consensus right now.

The DEFINE-RECORD-TYPE SRFI is the way it is because:

  1. I need some sort of record SRFI to make it possible to write
     other, more interesting, SRFIs.

  2. I have used a number of DEFINE-RECORD syntaxes, including
     ones with methods and ones with initializers, and I like this
     one the best.

  3. It is easy to understand and to implement.

It is absolutely not meant to be the be-all and end-all for records.

                                 -Richard Kelsey