Re: Scheme is too complicated

[Craig Brozefsky <craig@onshore.com>]

Jay Glascoe <jglascoe@jay.giss.nasa.gov> writes:

> yeah, for instance, the Scandinavians (?) are absolutely mad about Erlang. 
> comp.lang.functional FAQ:  "Erlang is a dynamically typed concurrent
> functional programming language for large industrial real-time systems."
> 
> but what I really want is a high-level (very or not ;) language for gluing
> my C extensions together, i.e. "Guile".

But since I'm assuming we all want a well developed scheme envrionment
that is not trapped by it's own homogeneity, not a bloated
parenthetical beast of a compromise with the world of
machine-specialized assemblers, we need to tread carefully along the
edges of scheme.

I think it's crucial that guile be capable of incorporating the rest
of the world into itself without crumbling.  I think this requires
careful design of the module system, and the interface to alien object
code.

CMUCL has a interesting alien interface, which allows you to describe
different values, and functions, all in lisp.  This means I can attach
to alien libraries without having to write any foreign code, or at most a
minimal amount.  I had a working, but incomplete ncurses interface in
about a night with it.  This seems congruent with Jim's idea of having
an ABI aware foreign data manipulation "library".  Here is an example
from the cmucl manual:

<startquote>
Alien types have a description language based on nested list structure.  For
example:
     
     struct foo {
         int a;
         struct foo *b[100];
     };

has the corresponding Alien type:
     
     (struct foo
       (a int)
       (b (array (* (struct foo)) 100)))
</endquote>

So you create Alien types, which are integrated with the rest of the
type system.

> I'm not sure I follow.  In python I can do
> 
> import math
> def area(r):
>     return math.pi pow(r, 2)
> 
> In Guile I would *like* to do, i.e
> 
> (uses-modules (core math))
> (define area (r)
>   (* math:pi (expt r 2)))
> 
> (and maybe I can?  I have no idea how Guile's module system works.)

Presently there is not support in the reader to handle that.  I
believe you can do it with the low level module, I think
(module-symbol-binding module symbol &opt value), but I havent tried
it myself.

Jim has said that he wants to support this type of name space.

> I'm sure some whiz-bang elisp hacker can whip up some special Guile mode
> for emacs.

I believe someone is, and basing it on ilisp, which is nice.