Julian Sanchez has an excellent post about the conceptual and practical oddities of the “obviousness” standard for patentability or non-patentability. But if you back up to an earlier step of his analysis I think it points the way toward a better standard: “Patents are not supposed to just be a pointless monopoly granted to the first person who happens to file a description of a particular invention with the Patent Office; the justification for granting the monopoly is that it (in theory) elicits innovations that would not exist but for the incentive a patent provides.”
It seems to me that a more rigorous idea about where it does and doesn’t make sense to grant time-limited monopolies is just to forget about the issue of how “obvious” something is and ask instead how capital intensive it is. If the only way to make something is to invest a ton of up-front capital in making it, then the monopoly grant serves as a way of attracting the capital. But if you’re talking about something that can be made with relatively little in the way of up-front capital, then you’re not so much incentivizing creation as you are raising the cost of innovation. This is what makes the burgeoning software patents category so pernicious. The pharmaceutical patent system is not without its problems, but the basic logic of “nobody would invest all this money in R&D if there weren’t windfall profits at the end of the rainbow” makes sense. That’s why proposal to change pharma patents always involve replacing the money with prizes or with direct public funding or something. But software’s not like that. Hiring a couple of programmers and buying them computers is trivially cheap compared to the equipment and manpower you need to develop a new chemical.
Talking about obviousness makes it seem as if the relevant issue is how much pure genius you need to pour into a new invention. But the relevant consideration here is actually financial incentives. You need large financial incentives in fields where innovation requires large up-front capital investments. When you don’t need the large investment, then what’s the incentive and incentive for?