One big reason for releasing the source code of our viewer has been to allow for innovation outside of Linden Lab, so it's great that McCabe and Jacek are starting this experiment. We know that our more careful approach to viewer enhancements of late has been frustrating for many, but a lot of residents seem very happy with the stability, performance, and UI improvements of the 1.20 release (McCabe's "Classic" skin was an essential component of this release). Going forward, having alternate viewers like Imprudence out there gives us greater latitude to remain careful about our release process, so that we don't need to throttle back on our quality assurance testing on our mainline product, while still letting others quickly iterate on future solutions. We're hoping that the UI innovation that McCabe and Jacek are working on will be work that will eventually find its way into the Second Life viewer.

@3 (JayR): we're slowly modularizing the system, trying to do so the least disruptive way possible. The bigger the rewrite, the longer it takes to get back up to the previous level of stability, so we have to be careful here. Moreover, as the Google Chrome release shows, there's still a debate about the right way to do plugins, where there's a pretty good argument to be made for greater isolation of plugins than exists in current day browsers. Committing to a plugin interface is a really big deal, and we think that it's better to wait and get it right rather than rush into something that commits us to an evolutionary dead-end.