Re: Byte-code compilation - good for loading.

[Travis Bemann <bemann at execpc dot com>]

Please note that I'm new to this mailing list.

Jim Blandy <jimb@red-bean.com> wrote on 20 Dec 1999 18:49:03 -0500:
> * I believe a bytecode interpreter could improve Guile's
> maintainability, by relegating the analysis of the source forms to an
> entirely separate phase from execution.  Each part of the process has
> simpler responsibilities, making it easier to write maintainable and
> efficient code.  (This would also fix Guile's current problems with
> macro memoization in multi-threaded code.  Marius and Mikael know what
> I'm referring to.)

I personally think that the division of code parsing and code execution would
definitely be a Good Thing.  This would essentially split Guile into two
portions - a code parsing section, and a code execution section.  It would be
easier to maintain them separately than have them intermeshed.  Changes to the
parser would not have a significant effect over the code execution section, and
changes to the code execution section would not have a significant effect over
the code parser section.

> * I believe a bytecode interpreter could provide much higher code
> density than the current representation.  If Guile is stretching the
> processor's data cache, this might improve speed, too.

I would personally assume that bytecode would be much denser than text code.
Also, the original code wouldn't have to be kept in memory during execution
with a bytecode interpreter.

> * I believe a bytecode could be a better target for multiple language
> support than Scheme.  I think common target languages should be very
> low-level and concrete, since this leaves more implementation choices
> in the hands of language-specific code, where they can be decided
> appropriately.

A bytecode interpreter would make it quite easy to totally convert Guile into
interpreting a different language, as long as the bytecode instruction set is
flexible enough.  One would just have to essentially put a different parser on
top of the bytecode interpreter, and small changes might have to be made to the
bytecode interpreter.

-- 
Travis Bemann