Specification of ratsubst WAS: Forcing substitution

ratsubst does indeed solve Owlett's problem.

But I'd be interested to know what exactly the *specification* of ratsubst
is.  Consider, for example:

    ratsubst( r, a+b, a*2+b ) => r+a
    ratsubst( r, a+b, a+b*2 ) => 2*r-a

Why the assymmetry?  Presumably something to do with variable ordering, but
I can't get it to act differently if I change ratvars or use CREs with
well-defined variable ordering, e.g. rat(a+b*2,a,b).

             -s