It's funny how emotive the patching strategy has become. Elsewhere in the world of development it's not uncommon to see a twin track approach to updates: 'frequent interim builds' and 'last stable build', with the former being aimed at activist-users (like the people here and in ST), and the latter aimed at users or organisations that are happy to trade [working ] features for stability. With ArmA2 we now have something akin to that, but without forcing us to make a choice: the slower 'official' patch track running from our main shortcuts, and the faster 'beta' patches from what is effectively a mod folder. So switching between builds is easy. Moreover, as many other posters in this thread have highlighted, the public nature of the new beta patches results in many more testing resources, which can only benefit BI's developers (and, eventually, all of us).

Stepping back, it seems easy to forget that both OFP and ArmA took a while to mature post-release, and yet both have provided this (wider) community with almost a decade of enjoyment. With its ambition and sheer scope, ArmA2 is shaping up to do much the same; it seems a worthwhile goal for us (collectively) to help BI iron out the bugs.

One other (back on topic) thought: with the current round of beta testing, would it be an idea to stop using any other mods in those tests? From a QA perspective it strikes me that using other mods at the same time introduces elements of variation that might distort the results and/or make it harder for the BI developers to reproduce the environment/steps-to-repeat.