I went ahead & changed nand & nor to be n-ary, as Marc suggested.

I also discovered a bug in the spec for n-ary eqv. This is true:
  (eqv i j) = (not (xor i j))
But it does not generalise to the n-ary case. That is, it is *NOT TRUE* that
  (eqv i ...)  =  (not (xor i ...))

This buggy n-ary definition appears in the spec and the implementation of
n-ary eqv. Here is the bogus definition from the reference implementation:

  (define (bitwise-eqv . args) (bitwise-not (apply bitwise-xor args)))

Nope. To use the not-xor definition, you must fold it across the args;
you can't save the not operation until the very end. So this is correct:

  (define (bitwise-eqv . args)
    (reduce (lambda (a b) (bitwise-not (bitwise-xor a b)))
	    -1 args))

or, sticking to the R5RS core:

(define (bitwise-eqv . args)
  (let lp ((args args) (ans -1))
    (if (pair? args)
        (lp (cdr args) (bitwise-not (bitwise-xor ans (car args))))
	ans)))

I have changed the SRFI accordingly. The new draft is at
    http://www.cc.gatech.edu/~shivers/srfi/srfi-33.txt
and will propagate out to the standard url
    http://srfi.schemers.org/srfi-33/srfi-33.html
in the hands of esteemed editor Solsona in the near future.

If anyone knows of a simpler n-ary definition of eqv that could
be simply expressed in terms of n-ary primitives like and, or and
xor, I would like to see it.
    -Olin