For some time I've felt that our Boost tooling could be improved. By boost tooling I mean things like b2, boost unit testing and integration testing, documentation generation, reports like dependencies, website, etc. Currently we rely on certain privileged authors to develop these things then then upload them to the boost directories. This has worked, but it could be better. We spend what seems to me an inordinate amount of time chasing down anomalies, researching errors, responding for requests for help, etc. I think design, robustness, documentation could be improved. I think that the the most natural way for us to do that would be to adopt the practices that we have used successfully for decades to develop superior quality libraries to the design and development of tooling. That is: a) Someone proposes a new tool or improvement in an existing one on this list. His proposal includes a working though likely incomplete prototype, documentation, tests, demo etc. b) This is subjected to the same formal review process that libraries are. c) Subject to this review, the proposal is either rejected, accepted or accepted with conditions. Of course this would only apply to new tools and updates of existing ones. Robert Ramey