The *business* of software

Court Rules Software Not Patentable – or Maybe It Is – And Do You Care?

On October 30 the US Court of Appeals for the Federal Circuit (CAFC), the “patent court,” issued a decision on In Re: Bilski. Because Bilski involved a business method patent, and because the court had announced that it would hear the Bilski case en banc – that is, with all twelve justices on the bench, as opposed to its use of three judges for routine hearings – the decision was widely-expected to be a landmark ruling with regard to software and business method patents. The ruling was indeed broad – but no one seems to agree on exactly what it means.

When most people think of patents, they think in terms of physical machines. But the law, the US Patent Act, says an inventor may receive a patent for “any new and useful process, machine, manufacture, or composition of matter.” In 1980, in Diamond v. Chakrabarty, a case involving a patent on bio-engineered bacteria, the Supreme Court looked at the Congressional committee notes and concluded that Congress intended that phrase to mean “anything new under the sun” with the exception of “laws of nature, physical phenomena, and abstract ideas.”

In the years following 1980 the CAFC initially took the position that business methods were abstract ideas or mathematical algorithms and therefore unpatentable because of Chakrabarty. However, in a series of decisions they gradually hedged that position when the algorithm was implemented on a computer and achieved a specific result.

The CAFC flung open the doors to business method patents in a 1998 ruling referred to as State Street. In it, they ruled in favor of a patent on a method for administering a pool of mutual fund assets. The State Street decision has been interpreted to mean that any process that achieves a useful result – even a method of exercising a cat – was patentable. State Street led to such patents as the widely-ridiculed Amazon patent on “one-click” checkout.

Business method patents jumped the shark when Bernard Bilski, the CEO of a small commodities firm called WeatherWise, and his associate Rand Warsaw attempted to patent a simple three-step method for using hedging to manage commodity price risk. Bilski freely admitted that the method did not require a computer for its implementation. The patent office turned the application down on the basis that “the invention is not implemented on a specific apparatus and merely manipulates [an] abstract idea and solves a purely mathematical problem without any limitation to a practical application.” Bilski appealed to the CAFC, and the CAFC announced the en banc hearing.

The case attracted thirty-five amicus briefs (arguments from interested persons not directly a party to the case) from organizations ranging from the ACLU (which argued that the Bilski patent would deny free speech because one could infringe it simply by discussing the hedging strategy) to IBM (which argued, not surprisingly, for the retention of a “computer” test in patents involving algorithms).

In the Bilski decision, the CAFC seems to say that in order to be patentable, an invention must either be tied to a specific physical device (such as a computer) or transform a physical object. The Bilski process was admittedly not tied to a machine, and as it did not transform a physical object, the CAFC held it to address unpatentable subject matter. The Bilski patent was rejected. (The court did not address the issues that thought processes are physical state transformations, or the physical nature of information in information theory.)

The reaction was immediate and noisy. Gene Quinn, on his Practicing Law Institute blog, wrote Federal Court Decides Software No Longer Patentable. Conde Nast Portfolio opined “The End of Business Process Patents.” Computerworld, playing on the day-before-Halloween date of the ruling, wrote “Boo! Scary software patents are dead, dead, DEAD!”

But after the weekend, cooler heads were starting to have second thoughts. “Software and business method patents may still be valid” wrote Dana Blankenhorn and Paula Rooney. A UK e-commerce law site wrote “US court narrows scope for business method patents.” The New York Times changed the headline on its Bilski story from “Federal Court Kills Patents on Business Methods” to “Federal Court Limits Patents on Business Methods.”

What will the impact be for the major software patent players like IBM and Microsoft? Any existing patents that do not meet the criteria of Bilski may not be valid. You can bet that any company that is paying royalties on software or business method patents, or that is threatened with an infringement action, is re-examining its strategies. Pending software and business method patents will have to be revised or may never be granted.

What about you? Do you agree with the Electronic Frontier Foundation that software patents should be severely limited? Or are you in a start-up that has used a software patent or application to demonstrate value to your investors? Will Bilski be appealed to the Supreme Court? What will they decide?

(Full disclosure: I have a software patent pending. I think people who invent with software deserve the same access to the patent system as people who invent with brass gears. I think there are some steps to make the software patent process more rational than it has been. Perhaps I’ll write more in the future about what I think some of those steps might be.)

Tags: bilski, business method, patent

Comment by Neil Davidson on November 7, 2008 at 10:00pm: Thanks for the clear summary of the case Charles.

In principle, I don't see any reason why somebody who spends a lot of time developing a non-obvious solution to a difficult software problem shouldn't merit the same protection as somebody who does the same for a physical problem.

In practice, however, the system is simply broken. If the system broken beyond repair, then rather than trying to fix it I think it should be abolished. Sure, this will hurt some people with genuine claims, but it will also help the majority - those at risk from submarine patents. The net effect would be overwhelmingly positive.

Comment by Charles Mills on November 7, 2008 at 10:26pm: Neil, thanks. I think there is a lot of misleading information circulating out there as "truth." "Submarine patents" basically went away in the US with the patent law revisions of 1995. Nearly all patents are now published openly on the Internet 18 months after filing, and last for 20 years from the date of filing, not the date of issuance, so there is little benefit to an inventor in postponing patent issuance.

Believe me, no one is more opposed to "submarine patents" than I am. Many years ago I read an article in Dr. Dobbs Journal about a clever compression technique "that should be in every programmer's toolbox." We implemented it in a commercial product that my company sold. Imagine my surprise when several years later Unisys popped up with a patent on the technique (LZW compression). We were faced with a very difficult re-engineering and "customer management" problem -- or the alternative of paying royalties to Unisys. We chose the latter path. I'm very happy that submarine patents went away thirteen years ago.

Comment by Bill Montgomery on December 1, 2008 at 5:04pm: In my mind the question should be, do obtaining patents really add value? The answer is unclear and in my experience, patents are seldom an important asset when it comes time to sell your company. Few small software companies can afford to defend them against a larger competitor and few buyers are willing to pay a significant premium to obtain them.

Comment by Perry Ismangil on December 15, 2008 at 10:53pm: As an open source enthusiast, I am against patents in general, software especially. So I am on the side of EFF and FSF on this one. The arguments for me are: does patents foster or hinders innovation? Both sides claim they foster innovations. I am on the side that they hinders innovation.

Comment by Charles Mills on December 16, 2008 at 5:47pm: Against patents in general? No one should have a right to profit from an invention? Or rather, they should have no more right to profit from an invention than the guy next door who makes a cheap copy? Do you really feel that way? Most inventions today are not "better mousetraps" -- they require expensive laboratories and/or teams of people and/or years of work. Why would anyone expend those resource -- why would any investor fund those resources -- if they could not recover those expenses because they would be undersold by the guy next door who had none of those expenses? Would not eliminating patents greatly discourage invention? Am I missing something?

Comment by Perry Ismangil on December 16, 2008 at 6:04pm: "Why would anyone expend those resource -- why would any investor fund those resources -- if they could not recover those expenses because they would be undersold by the guy next door who had none of those expenses?"

You're right. I don't know the answer to that. Maybe investors in Red Hat or other companies that do not have patents can answer better.

Other people far more knowledgeable than me have pointed out that innovation and invention happens even without patent protection. Or have also pointed out execution always wins even if everybody have the same idea.

Maybe if I am clever enough to invent something, I would have a different perspective :)

Comment by Charles Mills on December 16, 2008 at 6:24pm: I think you are running two things together when you mention Red Hat. LOTS of companies make money without patents. My company made money with no patents. But we were not inventors; we were implementers. We wrote a software product that solved a business need, but I would not call it an invention. The value was in the execution and the business we built around that.

But not all software is like that. I have invented a way of protecting software from piracy. I have not executed at all. I did my time running a software company and I have no desire to try to compete with Macrovision and Aladdin. One of the anti-patent arguments with regard to software is "what's wrong with copyright?" Copyright would have no value. Copyright protects the expression of ideas, not the underlying invention. Would you like to see the source code for my test program? I'd be happy to send it to you -- it has no value except as a test program. It's not the valuable thing; the underlying invention is the (possibly) valuable thing.

Now, granted, I am not a poster child for expenditure of resources. I had few expenses or investments, other than a modest amount of my time. But without some sort of protection for my invention, I would have done nothing whatsoever with it. The world would arguably be a little poorer. (Very little. So far, there has been no commercial uptake for my invention.) Absent a patent, and absent my willingness to start another software company to compete in a mature niche, what would be my incentive to put this conceivably valuable little technique out into the world?

Comment by Charles Mills on December 16, 2008 at 6:33pm: FWIW, I just searched the PTO's database. Red Hat appears to have SEVENTEEN issued patents and 27 pending applications.

Comment by Perry Ismangil on December 16, 2008 at 8:34pm: As you put it so well, I'm also an implementer not an inventor. Therefore I would not have any idea whether an inventor would, or would not be incentivised if he can't pick and choose the implementers.

As I've mentioned before this is always the crux of the argument: no patent = no incentive for inventors vs no patent = inventors keep inventing

I'm sure each camp got plenty of examples of each case.

Comment by Jon H. Williams on December 18, 2008 at 9:30pm: Were not the accumulation and protection of individual rights predicated on a mortality factor? While corporate interests experience no similar and unpredictable mortality, the patent system will remain crippled. I appreciate and would tend to support the EFF brief on this, particularly as I believe the purpose of the patent system has been lost or twisted. I would like to read what Dean Kamen has to say about this.