[<<][staapl][>>][..]
Sun Sep 8 00:21:01 EDT 2013

s-expression only?

So is it possible to define code as just an s-expression?
Currently I'm not sure.  It would have to be in the form

(define macro/foo (macro: 1 2 3))

And similar for variables and words, but not sure how it links to the
compiler.


Hmm... the Forth syntax is actually quite complex, i.e. quotations,
scheme forms, macros with arguments, ...


Rewrite is not an option.
( How many times did I go through this loop ? )

Structure is pretty much stuck.  Lots of features have support that is
quite intertwined..

See sexp-example.ss for an exploration of s-expression syntax.
I swear I did this before..

It shouldn't be too hard to define a special-purpose syntax for only
the DTF forth.  Requrements:

  - parsing words ':' and 'variable', '2variable'

Question is if this is going to be enough.  It looks like the
conclusion I got to last time (to include on-target compiler) has its
reasons: non-reflective prefix parsers just feel like a terrible hack.


So I guess good news and bad news.  Roadmap is clear, but what's not
clear where it will end up.  Maybe it is time for an s-expression
language?

The deja-vu is that I'm not happy with the base language, and the main
cause of that is the awkwardness of using pointers, and the fidgeting
with macros.  So it's clear: for real work, a 16-bit forth is necessary.

And to give it an edge, a huffman-threaded interpreter still sounds
like an interesting challenge.




[Reply][About]
[<<][staapl][>>][..]