The most notorious feature of the U.S. patent system is the current test for obviousness. On top of being "novel," patents are supposed to be "non-obvious." The novelty test is relatively straightforward: nobody must have previously described a device that has all the elements of the disputed patent claim. Thus, if A+B, B, and C have already been made or described, then A+B+C and B+C are novel but A+B is not.
In order to not bother people with trivial combinations, patent law has longed deemed that on top of novelty there be an "inventive step," which in the U.S. now goes under the name of "non-obviousness." To the Supreme Court in the 1950s, this meant that a patent claim must demonstrate "the extraordinary level of innovation, beyond the capabilities of a person having ordinary skill in the art, that warrants the award of a patent." This expressed most engineers' aspirations about what a true invention consists of, but it was notoriously subjective, leading to arbitrary decisions by patent examiners and courts. To create a more objective test, in the 1970's the Supreme Court held that a new combination which did not change the respective functions of at least some of the elements was obvious.
Then Congress created the Federal Circuit and granted it a virtual monopoly over patent appeals. The Federal Circuit proceeded to radically change patent law in a variety of areas, usually in favor of the patent seeker or holder. Until this most recent term, the Supreme Court has largely let it do so. But now, first with eBay and now with this obviousness case (KSR v. Teleflex), the Supreme Court is starting to rein it in.
Among the law the Federal Circuit changed, it made the obviousness test far more difficult for a patent examiner or alleged infringer to prove, but also made the test far more objective and straightforward. To be obvious, according to the Federal Circuit, there must be a teaching, suggestion, or motivation in prior literature that would have led a person of ordinary skill in the art (i.e. your typical engineer in the particular field) to have made the claimed combination. For the purposes of this counterfactual test it is assumed that this hypothetical engineer had all relevant prior art pasted on the walls of his office. The teaching or suggestion doesn't have to be a full-fledged description of a combination of the elements, as in novelty, but in practice the infringer almost has had to prove a lack of novelty in order to prove a lack of obviousness, rendering the obviousness test almost superfluous and leading to the raft of trivial patents engineers complain of these days.
I'm proud to say that the professors of two of the best patent classes I took in law school are leading the charge to overturn the Federal Circuit's obviousness test. John Duffy is co-counsel for the petitioner, KSR, which is challenging the Federal Circuit law. Robert Brauneis is Counsel of Record for the 24 Law Professors as friends of the Court who support reforming the obviousness test. These are two very talented individuals and I predict a splendid success.
KSR in its brief argues that the Supreme Court's test(s) should be restored, but it gives short shrift as to what specifically the test of obviousness should be. If you have good idea(s) for what the test(s) for obviousness should be, now is the time to let us know.
Patently Obvious has more discussion of the issue and links to all the various briefs.