The bottom line is that testing is IMO not working out, de facto becoming a bottleneck - and not only in terms of QA, but procedural difficulties. We have well over 100 packages in testing and there is no tendency of testing getting a better throughput, and there is almost no real solution for libraries. I would wish some council support on it (I would hate people see this as another community member self-initiatedly bossing them around). Of course I don't think *I* have to be that guy (Valerio headed some excellent QA oriented efforts), we might even have a team of QA 'moderators'.