Comments on: R7RS

By: James

James — Mon, 14 Sep 2009 19:23:51 +0000

One of the things that came up during the R6RS discussions IIRC was the ability to mutate imported symbols. People who wanted compilability were dead set against it, but people who want Scheme to remain compatible with SICP were in favor of it.

I think the difference between the products of Working Group 1 and Working Group 2 may not be that one contains fewer defined symbols than the other, but that one is more concerned with expressiveness and one with safety and the ability to reason about code.

By: James

James — Mon, 14 Sep 2009 18:09:40 +0000

... What I really want is a Scheme report that will unit[e] the Scheme community...

By: alaric

alaric — Mon, 07 Sep 2009 09:25:45 +0000

@Andy LeClair: Oh, PLT may well be a very nice implementation with modules and everything.

But it's just one implementation.

The problem with Scheme is that it's a family of languages; there's PLT, there's Chicken, and so on.

If we can bring the family closer together, so that libraries and even some applications can be shared between them where it's reasonable to do so, then we'll still have a family of languages - but without them all duplicating their effort in writing libraries.

If we're to gang up against Java, C++, PHP, Ruby, Perl, Python, and friends, we'll need to be united!

By: Andy LeClair

Andy LeClair — Sun, 06 Sep 2009 04:27:44 +0000

I think PLT Scheme solves all of the problems you proposed?

By: Andreas Rottmann

Andreas Rottmann — Fri, 04 Sep 2009 17:02:34 +0000

Regarding third-party libraries, I'm working on a "package manager" for R6RS implementations, called dorodango. It's main ideas are outlined in this thread: http://groups.google.com/group/ikarus-users/browse_thread/thread/fabb890e3015f6f1.

There's no public code yet, but it's not too far from being able to do something useful (for some value of "useful").