I should read mail before I send it. My point was not that environment
should create mutable environments, but rather that there appears to be
no way of creating a mutable environment for load, other than possibly
interaction-environment (whose mutability isn't defined). Accordingly,
both load and eval are claimed to be able to mutate a type of
environment that R7RS doesn't define officially (though
scheme-report-environment and null-environment only *may* immutable).

If the descriptions of the environment constructors are to stand as is,
I would like to see a sentence inserted explaining how an environment
may be created that can be used to load or eval a definition.
Alternatively, perhaps load and eval should be restricted to make
unspecified the use of environment specifiers when definitions are to be
processed.

-- vincent

_______________________________________________
Scheme-reports mailing list
Scheme-reports@scheme-reports.org
http://lists.scheme-reports.org/cgi-bin/mailman/listinfo/scheme-reports