proposing a simpler mechanism R. Kent Dybvig (13 Nov 2009 03:00 UTC)
Re: proposing a simpler mechanism Thomas Bushnell BSG (13 Nov 2009 04:23 UTC)
Re: proposing a simpler mechanism Taylor R Campbell (13 Nov 2009 04:31 UTC)
Re: proposing a simpler mechanism R. Kent Dybvig (13 Nov 2009 16:22 UTC)
Re: proposing a simpler mechanism Per Bothner (13 Nov 2009 16:56 UTC)
Re: proposing a simpler mechanism Thomas Bushnell BSG (13 Nov 2009 04:54 UTC)
Re: proposing a simpler mechanism Alex Queiroz (13 Nov 2009 13:44 UTC)
Re: proposing a simpler mechanism Marc Feeley (13 Nov 2009 14:24 UTC)
Re: proposing a simpler mechanism Thomas Bushnell BSG (13 Nov 2009 18:39 UTC)
Re: proposing a simpler mechanism Thomas Bushnell BSG (13 Nov 2009 18:36 UTC)
Re: proposing a simpler mechanism Alex Queiroz (13 Nov 2009 19:08 UTC)
Re: proposing a simpler mechanism Thomas Bushnell BSG (13 Nov 2009 19:21 UTC)
Re: proposing a simpler mechanism David Van Horn (13 Nov 2009 19:25 UTC)
Re: proposing a simpler mechanism Thomas Bushnell BSG (13 Nov 2009 19:36 UTC)
Re: proposing a simpler mechanism David Van Horn (13 Nov 2009 19:58 UTC)
Re: proposing a simpler mechanism Arthur A. Gleckler (13 Nov 2009 20:25 UTC)
Re: proposing a simpler mechanism David Van Horn (12 Jan 2010 18:51 UTC)

Re: proposing a simpler mechanism Thomas Bushnell BSG 13 Nov 2009 19:21 UTC

On Fri, 2009-11-13 at 19:01 +0000, Alex Queiroz wrote:
> Hallo,
>
> On 11/13/09, Thomas Bushnell BSG <xxxxxx@becket.net> wrote:
> > So what?  Are we now making srfi's have an inelegant interface because
> >  some implementations implement standard scheme poorly?
> >
>
>      Implementing integer sets is a bit difficult with "elegant" Scheme.

We don't need integer sets.  Scheme does not specify any lambda syntax
other than "this is the minimum number of parameters" and "there may be
extra parameters".  The only meaning I can understand for "arity" is
with reference to the formals lists in lambda expressions.  Some other
folks seem to have a rather more metaphysical understanding in mind, but
I can't quite figure out just what they mean by it.

In any case, multiple return values is a perfectly well defined part of
scheme, and it's easy to implement efficiently.  There are
implementations which can't be bothered, just as there are some which
can't be bothered to implement call/cc efficiently.  But that doesn't
mean we should start specifying language features in such a way as to
*require* an inelegant expression, because the elegant one is not
implemented efficiently by badly written implementations.

Thomas