Combining A+B and C may not be easy, but it is obvious. This is actually the main problem I see with software patents: idea C is "with a computer", and A+B is some existing invention. Newspapers - on a computer! Alarm clocks - on a computer! Bank transactions - on a computer! Sure it was hard to program them. It's still obvious. But if securing the bank transactions requires new innovations in security technology to glue the pieces together, those innovations could merit patent D. Does not and should not prevent anybody else from making their own secure bank transactions with a different security method because somebody got an A+B+C patent covering the obvious part.
Definitely, and that should be the answer to those:
"Alarm clock, on a computer!"
"That's obvious. Alarm clocks and computers both exist."
"But this was difficult because [intricate problem that's different with computer clocks] and we had to do [intricate solution]."
"Then put that in the patent claims."
Good patent examiners currently do that, but there's a bunch of terrible stuff out there.
Really not understanding your point about pharmaceuticals. How is the benzene ring different from "including a library or function in a program [which] should have an absolutely predictable result"?
Combine a program and a library and even before hitting compile, you should be able to tell exactly what the result is. Combine a benzene ring and a hydroxide compound and even if you done it at one position, move it someplace else and it could have the opposite effect. It's unpredictable.
I do agree though that pharmaceuticals are a bit different than other patent issues, but for a different reason: selling a drug requires round after round of expensive clinical trials because of the FDA. Without exclusivity, there may not be enough incentive for drug companies to pay for those trials if a generic manufacturer can reverse engineer the same drug and sell it on the cheap without paying for the trials. Maybe the FDA should have its own special exclusivity granting system so we can peel off one of the complications of patent law.
True. Pharmaceuticals don't really seem to mesh with patent law anyway - right now, a company will defend their patent application as I did above, saying that the result of any compound is absolutely unpredictable, so therefore, nothing is ever obvious in drugs... and then when they get the patent and some competitors makes a biosimilar drug, that first company will leap up and say it's just an obvious variation on the patent and is covered under the doctrine of equivalents.