Welcome to Software Patents

Pieter Hintjens

If you know me, you know that I take my work seriously [..] So imagine my shock [..] when I discovered first-hand that my lifelong profession -- building software products and convincing people to give me money in exchange -- was potentially illegal. Simply sitting down, taking a clear problem, and solving it step-by-step using my talent and experience could land me in court [..].

[I]n 2003 my firm received a polite letter from a patent lawyer representing a firm called AllisBlue, claiming that we were infringing on the "invention" of storing an email address and mobile phone number together in a database. We'd spent several years building a product, something like Twitter, that let people send and receive text messages in interesting ways. The economy was deep in crisis. We'd put most of our capital, maybe a quarter of a million Euro, into this product. The lawyer told us, "stop selling your product, and license our invention". They didn't sell software. Just the rights to use that trivial patent.

I called the European Patent Office (EPO) [..] and asked them, "how can someone patent software?" They answered, "that's not possible in Europe, it's forbidden by law". What? "Yes, the patent law specifically excludes data processing." [..]

I contacted the other Belgian wireless application firms, and over the next months we organized a forum to discuss the issue. Most of them had also been in touch with AllisBlue. One had already been in court. It turned out that AllisBlue consisted of, literally, one pending patent and a million Euro in cash. The money was just there to litigate [..]

It was scary to see how afraid [a lot of other] firms were of going to court. AllisBlue had an easy strategy. Rather than attack firms directly, it attacked their clients. It sued a newspaper for using an text message service. The newspaper canceled its service immediately. The firm providing the service turned around and licensed AllisBlue's patents. It never went to a real judge. The simple threat of going to court was more than enough to force these businessmen into unethical deals. They did the math, and it was cheaper to pay off the blackmailer than to stand up to him. AllisBlue didn't make software, so was immune to any counter-attack [..].

Patent law has several filters [supposedly a good thing] for "low-quality patents". (Oh, don't imagine for a second this is for your benefit as citizen. Patent law is written by the patent industry for the patent industry. These filters are just meant to stop general inflation of patents.) A patent must be new, it must be non-trivial, and it must not be for software [..] So one firm tried to oppose the AllisBlue patent by finding examples of existing software that did the same thing, years before. They failed because the EPO didn't accept that evidence. . Ah, you see, "prior art" is only valid if it's documented in a database the patent office will search. Which means certain academic journals, and previous patents. Since no-one had patented the AllisBlue "invention" before, it was, in the eyes of the EPO, "innovative".

Perhaps you're start to see how the patent system creatively redefines words like "invention" and "innovation", but I'm getting ahead of myself. Back to the story.

Our own product was barely making money. The mobile operators charged huge fees just to let us connect to their networks. I figured, if we pay one patent troll, what stops us being blackmailed by a second, and a third? I canceled the product, fired the team, wrote off the investment, and cursed the system [..]

[Then] I spent two years [..] focusing all my time and remaining money to slow down the patent system's expansion in Europe. It was an education: today I know more about patent and copyright law than even most specialized lawyers [..]

Let's explore some figures. I'm writing this in late 2011. In March, Microsoft paid Nokia about 1 billion dollars for a deal that gave Microsoft access to Nokia's patents. A little later, HTC bought S3 for 300 million dollars, to get access to S3's patents, mainly because Microsoft and Apple were attacking it. Then in July, a group of firms (Apple, Microsoft, Oracle, RIM, Sony, EMC, Ericsson) paid 4.5 billion dollars for a set of patents from Nortel. Google bid somewhat less, and lost the auction. In October, Sony paid 1.45 billion dollars for Ericsson's patent portfolio. Then in November, Google got Motorola Mobility, and its patents, for 12.5 billion dollars.

What's staggering here is that in 2011, firms paid about 20 billion dollars for patent rights. In 2010, Apple spent 1.8 billion dollars in R&D. Google spent about 4 billion dollars. The total for the whole software industry can't be much more than that 20 billion dollars.

Yet those patents are just for mobile phones, essentially yesterday's technology. My guesstimate of 20 billion dollars for R&D is mostly for cloud computing, social networks, and the rest of tomorrow's infrastructure. The patent wars for one segment of technology are costing as much as R&D into the future of all software technology combined.

One difficulty with tracking patent costs is that settlements are mostly secret. For instance, today Microsoft makes more money from extorting license fees from Android vendors than it does from its own failed mobile operating systems. Most of these deals are secret. Only a few firms, like Barnes & Noble, come out and fight [..]

I hope you're asking how this is possible, how a legal system can be so destructive and yet invulnerable. The answer is both simple, and profound. Put simply, the patent industry is so profitable that it corrupts anyone it does not destroy [.. And t]he irony is that without exception, a business that embraces the patent system will die. Motorola, Ericsson, and Nokia were once powerful firms that drove entire economies. Their slow death was predictable and I'll explain later how it happened, and why Microsoft and Apple will follow. It's such a predictable process that it deserves its own Law [..].

It's not just software. For decades, the [bad boys] of the patent system have been the pharmaceutical industry. The truth is that new drugs, like new software, are never the product of single teams working in shiny white laboratories. All innovation comes from mass effort, over time, to solve numerous individual problems. it's not just the formula for a drug. It's ten thousand answers to a thousand problems, along the whole process. At some stage these come together into viable products [..].

In 1980 the US passed the Bayh-Dole act [which] allowed universities to keep the patents on federally-funded research. This is an extraordinary model. The public pays for research that is then owned by private business and sold back to the public at an extortionate profit. Bayh-Dole was written and pushed by a consortium of universities and big pharma. Lucky universities privatized every significant part of the drug research pipeline, and sold these rights off to their big pharma partners. The public saw new drugs for HIV/AIDS, but at a thousand dollars a pop. Hey, that's what health insurance is for, right?

But what was really happening was worse than a simple hijacking of public research and blackmailing of the sick and vulnerable. The drug pipeline stretches 15-20 years. It's a slow process of turning data into knowledge. The process depends on collaboration between teams, across the world. Patents poison this process. You cannot at the same time collaborate on research, and file patents. Simply publishing or discussing a new idea means you can't patent it [..] Successful patenting demands secrecy, whereas successful innovation demands open discussion. Ah, you see the conundrum [..].

The drug research pipeline is layered. That is, a product floats on a sea of techniques and intermediate tools, and primary research, each a necessary step on the way to the market. As Bayh-Dole bit, and universities across the USA and then Europe and Asia started to cash in, they patented more and more of these layers, until even basic research techniques were patented as "products". Meaning that researchers could no longer share techniques and basic knowledge [..].

The consequences were predictable. By the late 90's, the drug pipeline was almost empty. Pfizer and Merck and the other big pharma brands saw their profits falling off a cliff [..] Today Pfizer's share price has fallen by 62% from its high in April 1999. The Bayh-Dole act has not cured cancer. It has not helped the drug industry past that 20-year pipeline window. The universities that jumped on that ship have lost their best researchers. It hasn't produced a single new success in medicine. Instead, it has allowed a few to profit from what was already in the pipeline, and poisoned the well of US medical research for at least three decades. If we're lucky, the research has gone elsewhere: to India, South Africa, China, Brazil. If we're unlucky, it has just gone.

Any firm that lives by patents will eventually die by patents [..].

[T]he move towards a digital economy means that ideas and inventions are no longer as tangible as they were. We once powered our economies with steam, but today it is with algorithms. There is no essential difference, it is one of scale. As ideas become more abstract, they become more layered. In the construction of steam engines there may be ten layers of ideas. In a web site, tens of thousands. Where a steam engine patent affected 100 firms, a software patent may affect a million businesses.