Re: No internal DEFINE-SYNTAX?

[William Clinger <will@skinner.cs.uoregon.edu>]

About half of the conceptual complexity of Jonathan's prototype
implementations of the high-level macro system were attributable
to internal definitions (and Jonathan's desire to write code that
would actually detect errors instead of behaving arbitrarily when
errors were made).  An internal DEFINE-SYNTAX would have increased
the number of error cases, and would have had a corresponding
increase on the complexity of implementations.

Any implementor that wants to allow internal DEFINE-SYNTAX is free
to do so.  I think anyone that actually implements internal
DEFINE-SYNTAX, while checking for the ambiguous cases, will
understand why we didn't make it part of the proposal.

Will