The extension of this idea to systems of equations is
quite desirable but non-trivial.
If we have k equations with k unknowns, then for each equation
we have many freedoms:
we can choose the variable on which to construct the iterator,
we can choose the iterator, and we can choose the order of the
And of course, an iterator for the system consists of k iterator equations,
one for each of the individual variables, so we can choose
each of these independently.
This means that we have a huge number of choices to build an iterator
for the entire system for small k, and the number of choices can
quickly become astronomical for larger k.
It would appear that this number is too high to justify its use.
We have some ideas which improve this situation,
but these are developed in a separate
paper, since they are out of
the scope of this one.