Fri Jul 31 12:37:40 CEST 2009

NEXT: study

I'm in a quite confused position right now.  Frankly I'm lost.  I need
to read up on some ideas that are key to the further evolution of
Staapl.  At this moment the current core is a dynamicly typed macro
system which has _names_ right: scheme's lexical scope + hygienic
macros.  This makes Staapl into an assembler with a fancy
preprocessor.  The next stage is semantics of stuff to build on top of
the untyped/dyntyped infrastructure.

The focus should be on _examples_ of how to tackle a particular
problem using constrained metaprogramming (typed DSLs).

1. bottom up semantics (ziggurat) - typed macros (dherman)

Figure out what the central idea is and try to see the connection with
writing DSL on top of the macro forth.

2. partial evaluation - abstract evaluation - staging - semantics


3. incorporate some special purpose static semantics (i.e. CSP/occam)