Oracle/Google: the patents and the implications

I don't follow how the author dismisses the security patent by judging one flowchart of an if/then statement, and then in the next sentence claims that the dalvik implementation is substantially different. If the patent is truly so simple, then the implementation can only differ by inverting the condition and swapping the if/then blocks. If the dalvik implementation is substantially different, then there's probably more to the patent than that one diagram.

From that point on, this analysis stopped smelling objective.

His point isn't that the patent doesn't cover Anrdoid, but that it's so broad that it should never have been granted.

That may be the case, but one can't establish that with a cynical joke about a drawing.

This "analysis" was awful all the way through.

This guy doesn't seem to know that Android includes Apache Harmony and not OpenJDK.

I think his theory of an implicit patent grant by virtue of GPL2 licensing sounds like a stretch and would need to be tested in court. If the FSF thought it were true there would have been no need for GPL3.