Wed May 28 12:43:27 CEST 2008

rethinking forth parsing

The problem to solve is to identify parser commands before code is
parsed. This is only possible when all s-expressions are collected
before linear source code is parsed. This means that it's not possible
to define parser words that expand to s-expressions.

There's a chicken and egg problem there that deserves some
attention. The problem is that identifiers can't really be identified
as in scheme. -> look into how this expansion stuff works.

This works fine in scheme:

     #lang scheme/base

     (define broem (bla))

     (define-syntax make-bla
       (syntax-rules ()
         ((_ bla)
          (define-syntax bla
            (syntax-rules ()
              ((_) (+ 1 2)))))))

     (make-bla bla)

So, just expand everything to this, should work fine.

Maybe the RPN compiler should be simplified a bit, using syntax
parameters instead of compile time parameters.. Makes sense?