Re: 2008?
That's no really a first-to-invent case... the patent was applied for in 2002, it just took them all this time to have it granted. First to invent lets another guy come along and claim the same patent, even if these guys had already filed. That's going away next year -- in January, we're first-to-file, just like the rest of the world.
That's actually a much better system, for the simple fact that if there are really that many different folks "inventing" the same thing, it must not be that unique. So someone files, and another guy comes along with proof they invented the patent first. That should invalidate the first guy on prior art, but the second guy also doesn't get the patent. Much better for everyone.
And yeah, the USPTO is crazy stupid about a bunch of things. They pretty much only search existing US patents for prior art. They don't mandate that examiners "skilled in the art" -- the specific area of the patent -- thus, they grant all kinds of drek that should fail the test of obviousness. Software patents don't even have to include source code anymore, so the actual implementation of the invention -- the thing that really matters (eg, you can patent a specific implementation of an algorithm, but in theory, you can't patent an algorithm) -- isn't included in the patent anymore.