[prodigal_erik]

"Obvious" in US patent law basically means you assembled pieces of prior art like Lego bricks. If you use something nobody ever wrote down, it doesn't matter how many of your peers would have immediately realized the same solution (i.e., whether your disclosure actually benefits anyone), you win the race just by having encountered the problem first.

http://en.wikipedia.org/wiki/Inventive_step_and_non-obviousn...

[magicalist]

Sort of. That's really part of the Federal Circuit's ever-expanding scope of patentability, which the Supreme Court has made some moves to hem in (as noted in the third paragraph in the section you cite), and is a somewhat recent development. The Supreme Court in fact ruled exactly the opposite to the Federal Circuit's test that you paraphrase, ruling that "obvious to try" is a valid obviousness test. In other words, even if encountering a totally novel problem, if a person with ordinary skill in the art would have also tried the same solution as yours, it should not be patentable. See http://en.wikipedia.org/wiki/KSR_v._Teleflex

The fact that the decision was unanimous has a lot of us hopeful that the new patent cases that the supreme court is taking on will see more of the federal circuit's nonsense cut back. It won't be patent reform, but it will be progress.