Proposal for EVAL

[Marc Feeley <feeley@IRO.UMontreal.CA>]

It seems to me that Jinx's proposal DOES contain the notion of "first-class"
environments.  The procedures NULL-ENVIRONMENT, INTERACTION-ENVIRONMENT,
and SCHEME-REPORT-ENVIRONMENT return a value that can be used like any
other value.  Granted, the proposal only specifies 3 possible environments
and no specific operations on environments other than EVAL.  A better
wording would probably be: "weak" first-class environments.  Anyway, this
is just a detail.

To better evaluate this proposal I think it would be good to have an
idea how it could be extended to "stronger" first-class environments
and how it interacts with other language features (current or
reasonable extensions).  For example, packages, macros and error
systems.  In particular, what are the MIT-Scheme extensions to this
proposal?  Also, what syntactic environment is used during evaluation?

I am not judging the proposal at this point (or Jinx's intensions :-),
I just want to have a better feel for the ramifications.  It would be
unfortunate if EVAL dirtied the rest of the language (and once again,
I'm not saying the current proposal does).

Marc