Would it also be useful to have an operation

(bitwise-= a ...)

in which bit i is set in the result if bit i is the same in all the
inputs?  You can't get this function by folding, since you have to
propagate three states for each bit: consensus 1, consensus 0, and no
consensus.

(I sort of expected eqv to behave this way, but then realized that
that doesn't fit in at all to Olin's agenda of providing names for all
the two-input boolean functions.)