Ron Jacobs is of the oppinion that the community should build their own application blocks. His proposal is quite interesting, because Microsoft would furnish test cases and do some design work and then let the community build the blocks themselves. The reason he gives is that the PAG could never really build all the blocks the want to build, so the community (who apparently has lots of extra time on its hands) could build them themselves.
Quite frankly, if I really wanted to code open source software for someone else all day, I would be using Java. I think this idea blows. If stuff is so difficult to code that you waste enough code to require an application block and so common that building an open application block is justified, why isn't it in the framework already (Take a look at ClickOnce vs. the Application Updater application block)? Admittedly, maybe Microsoft (who happens to be the most powerful software company in the world), just can't spare a few bucks to stick this stuff in the framework. However, why devote so much effort and expenses into coordinating a bunch of open source projects with high probablities of failure and code Microsoft will never really be able to integrate into the framework, when Microsoft could be spending that time making the framework better with a low probability of failure by using internal resources? I just don't get it. In my experience, the theory of open source software development is usually a lot better than the reality (perhaps Ron and the team have never owned any open source projects before?). The project owners must always do the bulk of the work, so this spec+tests thing just wont cut it. Don't believe me? Check out the contributor list for Mono and you will see that everything important has a Ximian guy's name all over it.