Given a stream s1 of constraints [A, B, C] and a stream s2 of constraints  [D, E]:

• a join() of both s1.join(s2) would give a stream of [ {A,D}, {A,E}, {B,D}, {B,E}, {C,D}, {C,E} ] which is not what you need.
• a concat(), union() or whatever we call it, doing s1.union(s2) would need to return [A, B, C, D, E]. This operation doesn't exist yet in the current ConstraintStreams API.