You answered your own question. We've already tried to make our own tests and failed, thereby significantly delaying our response. It's likely that we would fail to integrate the IP successfully on the first try too. Also what is best-in-breed? At this point, for maximum community benefit, it seems like we need cheap and fast to produce, not highest quality.