[R6RS] cost of letrec* semantics for internal definitions

[William D Clinger]

Kent wrote:

> By Version 7.0a, (Petite) Chez Scheme uses letrec* semantics for internal
> definitions and so for them will detect violations of the letrec*
> restriction.  It will detect violations of the letrec restriction when
> letrec is actually used.

I'm sure the implementors of several other systems
would say much the same thing.  What is clear, however,
is that many implementors do not think it is worthwhile
to enforce the R5RS letrec restriction, and their users
do not appear to be complaining too much about that.  I
conclude that the Scheme community has not demonstrated
much zeal in enforcing such restrictions for the mere
sake of enforcing portability/conformance to standards,
so portability/conformance to standards is not a good
argument for having the R6RS require enforcement of the
letrec* restriction (after it is properly formulated).

Will