While the differences between the Android fork and upstream Linux will introduce pain points during the merge, Linus takes the long view and rightly points out

While the differences between the Android fork and upstream Linux will introduce pain points during the merge, Linus takes the long view and rightly points out that the Linux project is poised to inherit a huge amount of code that has gone through rigorous real world testing and quite a bit of success. This is the great thing about the Linux project. If your work doesn't get accepted right away (someone may have objected to the lock model that Android uses, compared to the vanilla Linux lock model), well, just go out in the market and prove that your work is better. Typically with that kind of proof you can start to work on a way forward to integrate your work back in. It's a win-win for the project, since if your endeavor fails they're not stuck with it, but if it takes off they're poised to reap the rewards. In summary, Forks (that are used for experimental features, not political schisms) are GREAT!