On Sep 9, Brian Harvey wrote:
> > This obsession with a 50-page limit for the spec is suffocating
> > the language
>
> Are you proposing that there not be a WG1?
No, I didn't make any proposals.
On Sep 9, Brian Harvey wrote:
>
> 1. Eli said that we were spinning our wheels and just rehashing old
> battles, and that that was the fault of "purists."
No. Equating "purism" with "rehashing old battles" is as bogus as
making "beautiful" the opposite of "practical". I also don't consider
"small" to have better chances at being "pure".
> 2. I said that **on the contrary, I think we're making great
> progress** but that, since he wanted to draw a line in the sand, I
> would be proud to be on the other side of it, **if we weren't able
> to make progress**.
I also didn't draw any lines. I only ranted about re-opening issues
that were thoroughly discussed, and ignoring some of these
discussions.
> As I understand it, this is why there are WG1 and WG2 in the first
> place; the SC recognized that both people like Eli and people like
> me are legitimate members of the Scheme community, with legitimate
> interests to pursue in the standardization process. Eli seemed to
> want to write me out of the community, and all I was saying was that
> I'm proud to be a WG1-Scheme user rather than a WG2-Scheme user like
> him. Not that there shouldn't be WG2-Scheme. Just that we
> WG1-Scheme people aren't blights on the community.
* I have absolutely no desire to write anyone out of any community.
(Related to this, I do not try to define "my community", since my
definition of any such thing would give me power over it, like
deciding that "my community" is "everyone except for Foo".)
* I dislike this WG1-WG2 business because (as said many times, here
and elsewhere) this division looks to me like an artificial one. I
can't think of a single schemer who is "clearly on WG<N>'s side" for
any N. Specifically, I don't consider myself a "WG2-user". In
fact, I was absolutely serious about the lambda calculus: if you're
looking for a "pure Scheme core", then LC + a few primitive types is
much closer to it -- and in that sense R4RS *is* still bloated.
* It looks like since you've pegged me as a WG2-user, you somehow
assume that I'll want the kitchen sink in each and every language I
use and/or talk about. That's very wrong. In my course, one of the
(many) languages that we use is a language that I introduce as
"almost like scheme, with a few differences". The differences are:
(a) it is lazy, (b) it has only one-argument functions and
applications, but the syntax allows more arguments by being a
syntactic sugar for the curried forms, (c) `define' is only a
toplevel form and it doesn't have the magic power of self-reference
(each definition can only use previous definitions), so you get
recursion only using Y, (d) other than identifiers, lambda
expressions, and function applications the language has nothing else
(things like `17' are also identifiers). I then go on to show how
it is possible to get the usual things using these basic tools.
The resulting language makes R4RS look as long as the lord of the
rings.
To implement it, I didn't write an evaluator -- it's all just plain
(PLT) scheme, so it's running relatively fast. It would be possible
to actually use this language in practical situations where you want
a restricted combinator language. It would be possible (and in
several ways much easier) to have a similarly restricted language
that is closer to Scheme -- with the usual semantics for `define',
`lambda', and applications, and with a few primitive types like
numbers and symbols.
But of course such languages are possible because I'm building on
PLT which has a lot more than r4rs. So according to the artifical
WG{1,2} division, I'm very much on both sides.
--
((lambda (x) (x x)) (lambda (x) (x x))) Eli Barzilay:
http://barzilay.org/ Maze is Life!
_______________________________________________
r6rs-discuss mailing list
[email protected]
http://lists.r6rs.org/cgi-bin/mailman/listinfo/r6rs-discuss
