An efficient and portable definition of `equal?' that does what you
would expect (tests for structural equality) in the case of
random-access pairs and lists can be defined in terms of the procedures
provided by SRFI 101, however the implementation requirements imposed by
R6RS for `equal?' are quite sophisticated.  For this reason, I propose
adding `equal?' to the specification and reference implementation to
avoid duplication of efforts.

I propose the behavior of this procedure should be unspecified on
systems with disjoint linear-access and random-access pairs when given
two pairs with differing representations.

Unless there are objections, I will add this to the next revision of the
SRFI.

David