Two thoughts on drupal.org stats: 1. I suspect our relatively disappointing stats are not unrelated to recent discussions on what content we present to whom. I'm thinking in particular of the definable group "potential new drupal users", which will include most firt-time drupal.org visitors as well as some repeat visitors. Say we have fifteen pages that answer the needs of this main group. Their needs - and therefore their content, ideally - should need to change very little--every few weeks, I'd say. As things stand, pretty well every page contains content that changes hourly (the "active forum topics" block)--and that is of little interest to potential new users. True, we could address this partly by caching page parts--but it would be much better not to show those parts to uninterested users in the first place. Also, it's apparently the case that few visits are from authenticated users. But if authenticating gave you a whole new view of the site, catered to your interests and needs, likely more of our registered users would take the step of authenticating. In short--doing a better job of presenting focused content to user groups should have performance benefits as well as other payoffs. Coding to make this possible would start with user group specific filtering of content views (blocks, home page, etc.). 2. Do we do any automatic revisiting/recaching of recently viewed/newly added/popular content after cache flushes? Might doing so help to preempt many dynamic page loads?