Re: `undefine'

[Mikael Djurfeldt <mdj@nada.kth.se>]

"Marisha Ray & Neil Jerram" <mpriz@dircon.co.uk> writes:

> >OK, just to make it clear:  The docs I posted are *old*.  While some
> >things might still be valid, most aren't.  I posted it only for the
> >discussion about unmemoization.
> 
> I'm a little confused.
> 
> A large part of your docs is very similar to Jim Blandy's recent
> environments proposal - presumably the two are related.  Is it correct that
> Jim Blandy's proposal is the more recent, and that "environment" is now the
> preferred name for "space", or is it the other way round?

What I posted is old.  Jim Blandy's environment proposal is new.

Yes, they are related, but the environments are even more clean than
what I posted (which is older ideas from Jim).

The idea is that module system funcionality can be implemented in
"layers" with clean interfaces in-between.  E.g., unmemoization would
be a property of the module system code.  This, in turn, uses the
observers of the environments code.