and the plan currently is for the resolves to be named, but for their inputs to be computed based on which roots are using them.