Why the Idea of the Software Patent Does Not Fly

A couple of weeks ago, I wrote a post about the growing threat of software patents in Europe (again). I was delighted to come across this reply from Martin Goetz, who wrote: On April 22nd I read Glyn Moody's Computerworlds.UK blog Software...


A couple of weeks ago, I wrote a post about the growing threat of software patents in Europe (again). I was delighted to come across this reply from Martin Goetz, who wrote:

On April 22nd I read Glyn Moody's Computerworlds.UK blog Software Patents Storming Up the Agenda Again and several thoughts immediately struck me. For one, he was factually wrong on several fronts. Secondly, it reminded me of an article I just published in the US that discussed machine translation. That's the main argument made by Mr. Moody that the European Inventor award should not be given to a group that invented a phrase-based machine translation using a statistical approach. I also was struck by his comment "The inclusion of a pure software patent (the machine translation patent) in the European Inventor Award shortlist is a real slap in the face of European companies and citizens, and looks like a calculated provocation from the EPO."

So, where did I go wrong?

Mr. Moody's factual mistake is that he believes that copyrights are a substitute for patents. Patents protect the invention while a copyright only protects the illegal copying of computer programs. A copyrighted program is not able to protect any invention (or "idea", as Mr. Moody calls it) that might be embedded in that computer program. It's just not a viable substitute for patent protection. That's why there are many thousands of software patents filed in the US each year.

What Mr Goetz calls a bug, I call a feature. One of the key problems with software patents is that they give people a monopoly on a programming idea: that's as absurd as giving someone the monopoly on the idea of a mousetrap. What patents should do is stimulate the creation of better, different mousetraps. That's precisely what copyright does in the context of software: it protects the details of that nifty subroutine that you wrote to implement a general idea, but does not stop anyone else coming up with a different approach. If somebody is able to patent the underlying idea of the subroutine, that clearly locks out all the better implementations of it, which means that patents actually block innovation.

So let's move on to Mr Goetz's article that he mentions. Here's the key argument:

Imagine that the digital computer and the stored computer program (software) existed in the late 1800's. If so, there is a high probability that the Wright Brothers would have used a computer program to control the three-axis control system in their 1903 Flying Machine patent application. If they did, we would call the patent they received in 1906 a software patent.

Further imagine that today's anti-patent zealots who preach that all software consists of abstract ideas, mental processes or mathematics were thrown back in time to the late 1800's. And finally imagine that they were successful in their stated goal to eliminate all software patents through an act of Congress. Had that really happened, the Wright Brothers would not have received a patent for what is recognized as one of the greatest inventions of the 20th century.

So the logic here seems to be that had there been software back then, the Wright brothers would not have received a patent on their software-implemented invention, and that the world would have been the poorer as a result. Well, let's just examine that example a little more closely.

Here's what Wikipedia writes about the Wright brothers' patent:

In 1906 the Wrights received a patent for their method of flight control which they fiercely defended for years afterward, suing foreign and domestic aviators and companies, especially another U.S. aviation pioneer, Glenn Curtiss, in an attempt to collect licensing fees. Their legal threats suppressed development of the U.S. aviation industry for several years.

One of the references for that article has more details:

The Wrights further restricted aviation progress in the United States by sticking doggedly to their basic design, despite the obvious advances being made in Europe. Improvements were made to the 1910 Model B, which had the elevator in the rear, wheels in place of skids, and did not require the tower-catapult for takeoff. The later Model C proved to be a man-killer; seven were purchased by the Army and five crashed, killing five men.

Thus it seems that the Wright brothers' obsession with their patent on "one of the greatest inventions of the 20th century" caused the US to fall behind Europe in aviation technology by failing to incorporate the advances being made on that continent, which largely ignored the patent. Indeed, the damage caused by the Wright Brothers' patent was so great that the US government was forced to intervene:

In 1917, the two major patent holders, the Wright Company and the Curtiss Company, had effectively blocked the building of new airplanes, which were desperately needed as the United States was entering World War I. The U.S. government, as a result of a recommendation of a committee formed by Franklin D. Roosevelt, then Assistant Secretary of the Navy, pressured the industry to form a cross-licensing organization (in other terms a Patent pool), the Manufacturer's Aircraft Association.

So it turns out that Mr Goetz's example is, in fact, the perfect demonstration of why patents are harmful. His hypothetical early 20th-century world where software existed, but software patents were forbidden, would have been spared the enormous problems caused by the Wright brothers' obsession with defending patents instead of coming up with better and safer products. The US government would not have needed to intervene to cut through the patent thicket, and maybe those unfortunate individuals would not have been killed by the Wright's outdated Model C "man-killer."

Some might say that without the patent, the aircraft industry would never have flourished, but a moment's thought shows why that's nonsense: the utility and value of machines that could fly is so high that people would have kept on developing them regardless of whether they could take out patents. That's because the real money is made by selling planes, not by licensing the technology that they use (think of Boeing and Airbus.) In the absence of patents, the main difference would have been a more rapid pace of development, as each company was forced to keep innovating to stay ahead of its competitors, spending money on research and development, not lawyers, as the Wright brothers did.

But let's move on, and look at two more of Mr Goetz's arguments in favour of software patents. The first concerns inventions that mix software and hardware:

In many industries inventions contain software programs as part of their disclosures. Take the Robotics/Medical Devices sector, where very sophisticated computer programs can control an artificial limb. How the artificial limbs are designed and constructed may be the invention. Or the invention may be how the computer program(s) interacts with the artificial limb. Or the invention may be a combination of both elements.

A more recent real life invention is the artificial retina, a robotics/medical device which offers partial vision for the blind. It's described as a camera, transmitter, and cable to video processor (software) connected to the brain.

So those that want to eliminate software patents, by design or by ignorance, would eliminate all robotic/medical device inventions where the disclosure includes a computer program.

There are two parts to these inventions: the hardware and the software. The hardware could be patented if it contained some new, non-obvious innovation. The software would be protected by copyright automatically. Again, that's as it should be, since granting a patent monopoly on an entire class of software implementations simply ensures that better ways of realising the underlying idea are suppressed, as discussed above, which is not at all beneficial for society.

So far from "eliminating" all such robotic/medical device inventions, excluding software patents here would in fact stimulate more competition, faster innovation and better, cheaper products – just as Europe innovated faster in aviation without the deadening constraint imposed in the US by the Wright brothers a hundred years ago.

Mr Goetz's final category is the "pure software" invention. Here's what he writes:

Anti-Software patent zealots might argue that the definition of a "software patent" is when the disclosure of the invention only describes a computer program and a computer. And then their goal would be to have Congress change the Patent System to eliminate the issuance of "software only patents". But that would be catastrophic for the Software Industry, where many of their inventions are software only. This industry is made up of thousands of companies and is recognized as one of the top three manufacturing industries in the world.With annual worldwide revenues well over $300 Billion this industry needs (and wants) just as much patent protection as other industries.

But patent protection is a very recent feature of the software industry which grew up and thrived without them. Indeed, early pioneers recognised that had software patents been widely granted, the software industry might never have taken off. Here's the view of one person you might have heard of:

In a memo to his senior executives, Bill Gates wrote, "If people had understood how patents would be granted when most of today's ideas were invented, and had taken out patents, the industry would be at a complete standstill today." Mr. Gates worried that "some large company will patent some obvious thing" and use the patent to "take as much of our profits as they want."

Yes, in one of his more candid moments, Bill Gates admitted – in writing – that software patents would have destroyed his industry. And as a paper by James Bessen points out, this opposition to software patents was widespread in the industry in the 1990s:

In 1994, the Court of Appeals for the Federal Circuit decided in In re Alappat that an invention that had a novel software algorithm combined with a trivial physical step was eligible for patent protection. This ruling opened the way for a large scale increase in patenting of software. Alappat and his fellow inventors were granted patent 5,440,676, the patent at issue in the appeal, in 1995. That patent expired in 2008. In other words, we have now experienced a full generation of software patents.

The Alappat decision was controversial, not least because the software industry had been highly innovative without patent protection. In fact, there had long been industry opposition to patenting software. Since the 1960s, computer companies opposed patents on software, first, in their input to a report by a presidential commission in 1966 and then in amici briefs to the Supreme Court in Gottschalk v. Benson in 1972 (they later changed their views). Major software firms opposed software patents through the mid-1990s. Perhaps more surprising, software developers themselves have mostly been opposed to patents on software.

Those original fears about the harm that software patents would cause to the industry have turned out to be well founded. The book "Patent Failure" by Bessen and Meurer has calculated the aggregate annual patent profits and costs for the years 1996 to 1999, shortly after software patents became more common. The annual patent profits for US companies from software patents was $100 million; the annual litigation costs for software patents were $3,880 million. That means on average, the net effect of software patents was a loss of over $3.5 billion per year.

Moreover, things have got worse, not better, in recent years, thanks to the rise of NPEs – better known as patent trolls – that have exploited patent monopolies being given for general ideas rather than specific implementations, as discussed above, to sue very broadly:

Firms that license patents without producing goods—" non-practicing entities" (NPEs)— have historically facilitated technology markets and increased the profits that small inventors earn from their inventions.

But a self-described new crop of NPEs has emerged that asserts patents and litigates them on an unprecedented scale, involving thousands of defendants every year in hundreds of lawsuits. Do these litigating NPEs improve markets for technology and increase incentives for small inventors? Or are they "patent trolls" who exploit weaknesses in the patent system?

This paper makes several findings about this litigation. First, by observing what happens
to a defendant's stock price around the filing of a patent lawsuit, we are able to assess the effect of the lawsuit on the firm's wealth, after taking into account general market trends and random factors affecting the individual stock. We find that NPE lawsuits are associated with half a trillion dollars of lost wealth to defendants from 1990 through 2010. During the last four years the lost wealth has averaged over $80 billion per year. These defendants are mostly technology companies who invest heavily in R&D. To the extent that this litigation represents an unavoidable business cost to technology developers, it reduces the profits that these firms make on their technology investments. That is, these lawsuits substantially reduce their incentives to innovate.

Patent trolls reduce the incentives to innovate in one field above all:

the characteristics of this litigation are distinctive: it is focused on software and related technologies, it targets firms that have already developed technology, and most of these lawsuits involve multiple large companies as defendants.

Just as with the Wright Brothers, then, patents are not helping innovation, but hindering it severely, particularly in the field of software, where hundreds of billions of dollars of wealth has been destroyed by patents. The overall effect on research and development is devastating:

In the smartphone industry alone, according to a Stanford University analysis, as much as $20 billion was spent on patent litigation and patent purchases in the last two years — an amount equal to eight Mars rover missions. Last year, for the first time, spending by Apple and Google on patent lawsuits and unusually big-dollar patent purchases exceeded spending on research and development of new products, according to public filings.

The evidence is overwhelming: software patents were never needed to create the software industry we know today, and they are no longer needed now that unpatented open source is replacing proprietary software at every level. Software patents do not promote innovation, they stifle it. They do not reward inventors, but patent trolls. As an idea and in practice, the software patent – rather like the Wright brothers' lethally-dangerous Model C aeroplane – simply does not fly.

Follow me @glynmoody on Twitter or identi.ca, and on Google+

"Recommended For You"

Software patents harm both industry and consumers but SMBs suffer more Cisco, SAS, Adobe tell US Congress to target patent trolls