This seems to be a large change at a critical time for SD work. I think that many of the gains are to be realized in synthesis so I would propose that we implement this first in synthesis and then later in SD. After trying an implementation in synthesis, we will issue a modified proposal based upon that experience. At the time of writing the second revision of this proposal, Mark Wieringa has made excellent progress in developing an implementation of the proposed scheme and expects to finish within a week.