Issue: UNDEFINED-VARIABLES-AND-FUNCTIONS

[Kent M Pitman <KMP@STONY-BROOK.SCRC.Symbolics.COM>]

Deferred until next meeting.

There was some ``off-line'' discussion of slot-unbound with no real resolution.
That part of the issue promises to be a controversial one.

Moon seems to think that one idea that had some support is to say that the
system-defined method "should signal an error", depending on the safety level
of the original call to SLOT-VALUE, so the generic function is only called if
the SLOT-VALUE is safe or user-defined methods might be applicable. I haven't
had time to decide whether I understand this enough to either agree or disagree.