[Disclaimer: this is a rather specialized question]
It is known that techniques like Conflict-Driven Clause Learning (CDCL) and back-jumping -- which improved the Satisfiability (SAT) strategies exponentially -- were actually inspired  by Constraint Satisfaction Problem (CSP) techniques (so-called nogoods , or generalized nogoods).
But while SAT evolved and improved with these techniques, CSP doesn't seem to widely benefit from these ideas.
My question : why aren't these ideas ubiquitous in CSP like they are in SAT ? What happened ? What makes them unsuitable ? Is it "just" a question of domain size ?
Secondary question: are there any other ideas that jumped the divide between the two ? Instances of the CSP community inspiring the SAT community and/or vice versa ?
: Generalized NoGoods in CSPs, G. Katsirelos, F. Bacchus
: Using CSP Look-Back Techniques to Solve Real-World SAT Instances, R. Bayardo, R. Schrag