Marc:
Thanks again.
I've included some suggested changes for _num.scm. They change "first-set-bit" to "first-bit-set" consistently; they use the fixnum routines ##fxbit-count, ##fx-length, and ##fxfirst-bit-set in ##bit- count, ##integer-length, and ##first-bit-set. The changes to ##bit- count, ##integer-length, and ##first-bit-set have been tested rather thoroughly, and I think the name changes in the rest are fairly straightforward to check.
I've also removed one optional clause in the computation of bignum multiplication. That clause just substituted one O(N\log N) algorithm for another and I want to simplify bignum multiplication in the short term.
Brad