issue ARRAY-TYPE-ELEMENT-TYPE-SEMANTICS

[sandra%defun@cs.utah.edu (Sandra J Loosemore)]

I don't see anything obviously wrong with the general approach taken
by this proposal.  However, I think the new functions added by
sections 7 and 8 are unnecessary and I would either like those parts
removed, or something added to the "Rationale" section to explain why
some people think they *are* necessary.  Also, since this proposal
changes the behavior of TYPEP on array type specifiers, I'd like to
see some clarification of how functions such as BIT-VECTOR-P that are
defined in terms of TYPEP in CLtL are affected by the change, or an
explicit statement that they're not.

-Sandra
-------