Samsung and Apple attorneys battle over software patents, Nov. 15. Reuters

Until the court changed the rules, there hadn’t been patents for algorithms and software. Ideas alone aren’t supposed to be patentable. In a case last year involving medical tests, the U.S. Supreme Court observed that neither Archimedes nor Einstein could have patented their theories.

Many software patents simply describe ideas that happen to be carried out through digital technology. Amazon got a patent for the concept of “one-click checkout.” Apple last year applied for a patent on the idea of offering author autographs for e-books. There are so many software patents that smartphones include some 250,000 purportedly patented processes, which is why Google,Samsung and Apple are suing one another around the world.

In software, innovations build on one another so seamlessly there is no way to follow them. There is no national registry of software. Developers and engineers can’t track who claims patents to what processes. In contrast, drug researchers consult a publication called the Orange Book that lists all the patents for pharmaceuticals, enabling them to avoid infringements.

A system of property rights is flawed if no one can know what’s protected. That’s what happens when the government grants 20-year patents for vague software ideas in exchange for making the innovation public. In a recent academic paper, George Mason researchers Eli Dourado and Alex Tabarrok argued that the system of “broad and fuzzy” software patents “reduces the potency of search and defeats one of the key arguments for patents, the dissemination of information about innovation.”

The Government Accountability Office agrees. “Many recent patent infringement lawsuits are related to the prevalence of low-quality patents; that is, patents with unclear property rights, overly broad claims, or both,” it said in a recent report. “Claims in software-related patents are often overly broad, unclear or both.” Boston University law professors Michael Meurer and James Bessen have estimated the direct and indirect costs of litigation against technology companies at $80 billion per year.

Instead of focusing on the problem with software patents, reforms backed by the White House and Congress would tweak patent litigation for all industries. The House this month passed a bill requiring more specificity in claims and limiting costly discovery, but doing nothing about dubious software patents.

The House rejected a proposal that would have expedited the process for the Patent Office to review questionable software patents. Lobbyists from companies like IBM and Microsoft, which make billions of dollars a year from licensing software patents, helped block this reform.

For now, the best prospect for real reform is in the Supreme Court, which earlier this month agreed to hear CLS Bank v. Alice Corp., a case about whether a bank’s computerized process for settling transactions via an escrow can be patented. A judge on the appeals court noted this idea was “literally ancient,” developed during the Roman Empire, and should not get a patent now just because a computer is involved.

The Supreme Court has invalidated software patents in earlier cases, but the justices need to draw a brighter line with clear limits for the lower courts, especially the Federal Circuit. Simply qualifying ideas or business processes with the phrase “and do it on a computer” shouldn’t be enough.

The justices should also acknowledge that creating a special court to promote patents is an experiment gone awry. Far from helping the economy, software patents are a litigation tax on new technology. The Constitution calls for patents “to Promote the progress of Science,” not for patents to undermine innovation.