Wed Jul 30 08:51:04 CEST 2008

Onwards: concurrency / types / ans

* This is interesting:

  I'm a bit sad alloted time didn't allow to implement the distributed
  debugging system for KRIkit last year.  But surely, concurrency is
  the next step for Staapl, next to some more static analysis.

* Reading about types lately, especially TAPL.  This article
  summarizes it well:

  Types are "things to prove about programs".  This can really mean
  anything.  According to Chris that's the big idea. This is
  significantly different from "types are sets" in the dynamic/lisp

* ANS Forth on top of Staapl Forth?  It is not straightforward but
  could take the form of a standard Forth compiled to staapl
  primitives with simulated dictionary.  More importantly, is it
  necessary?  Should Staapl contain a mechanism to build a standard
  reflective Forth on top of the unrolled macro Forth?  It looks like
  this is a ``marketing problem'' more than anything else.