Today, the Supreme Court hears arguments in Alice Corporation Pty. Ltd. v. CLS Bank International, a case that could massively overhaul the rules for software patents.

First, a brief patent primer. Adding one number to another is blatantly unpatentable, but a specific calculator that carries out addition (and a good deal more) is eligible for a patent. The question at the heart of Alice v CLS is whether a company could patent the idea of writing a calculator program, provided no one else had staked a prior claim. Under this system, whether or not Alice Corp built a calculator or even wrote the code might be irrelevant. As long as they were the first to think of using software to carry out this everyday operation, they would retain an intellectual property claim.

Alice Corp didn’t patent something quite as old as addition, but what they did patent—escrow, the practice of reducing the risk of a transaction by having a third party hold the goods to be exchanged until both sides have paid up—considerably predates Alice Corp’s 1993 patent filing. Alice Corp didn’t try to patent the idea of escrow, but the idea of managing it through a computer. A general description of the idea was all that was needed to file the patent; Alice Corp wrote no code and built no prototype.

In 2002, CLS built a working program to manage escrow-secured transactions in foreign currency markets, and Alice Corp, which had never put its patents into practice, sued for infringement. The case has been working through the courts ever since. When it reached the U.S. Court of Appeals for the Federal Circuit, it produced an exceptionally messy result. The ten judges hearing the case en banc produced seven different opinions.

The case law becomes very complicated, very fast, as, in order to decide if Alice Corp is patenting an “abstract idea” as CLS claims, the justices need to wade into very abstract questions. Alice’s lawyers contend that, by the logic of Mayo v. Prometheus the only kind of ideas that are too abstract to patent are “those preexisting fundamental truths, such as mathematical formulas, that are ‘equivalent’ to a law of nature and that ‘exist in principle apart from any human action.’”

By this logic, escrow may be a universal idea, but it’s not built into the nature of the physical world, so it’s fair game for patents. Certifying such a broad range of ideas open to patent could legitimize the actions of patent trolls, who file claims to a broad range of ideas they have no intention to develop, in order to shake down companies for settlements.

On the other hand, striking down Alice Corp’s claim might open the patent system to other forms of gamesmanship and abuse. If the courts take a very broad view of “laws of nature” most software patents could be on the chopping block. After all, who’s to say that Match.com’s dating algorithms aren’t simply a software implementation of certain universal heuristics for compatibility?

And a very limited ruling may just pave the way for more snarled rulings like the en banc hearing, which revealed the legal uncertainty that could deter companies from entering markets covered by ambiguous patents.

The opinion of the justices will need to thread a very delicate needle if they wish to keep software patents in existence but resilient in the face of abuse. And, unlike Alice Corp, the Supreme Court won’t be able to get away with simply asserting that a solution to this abstract problem exists.