Sunday, November 09, 2008

The practical problem with software patents

Note: An edited version of this article recently appeared in Economic Times.

The Bilski case judgment has reversed the trend of granting patents to abstract ideas in the US, and is good for software developers, says Venkatesh Hariharan

In their book, “Patent Failure: How Judges, Bureaucrats, and Lawyers Put Innovators at Risk,” Boston University professors, James Bessen & Michael J. Meurer, show that Murphy's Law (“If anything can go wrong, it will”) has been working overtime in the area of software. The authors dedicate an entire chapter to software and business method patents, which are particularly problematic because they account for almost 38 percent of all patent litigation.

The authors find that in the United States, software patents are twice as likely to be litigated as other patents while business method patents (which act as a proxy for software patents) are seven times as likely to be litigated. The authors say, “Our reading of the case law convinces us that patent law tolerates too many software claims untethered to any real invention or structure; in such a world clear boundaries are unattainable.” The authors point out that patent on abstract ideas are often subject to multiple interpretations and are therefore more ambiguous. An example of this ambiguity is the E-Data patent on "point of sale location." In the IT industry, this term is jargon for the cash register or location where the customer pays the cashier. When the US Federal Circuit interpreted this claim, they decided that it referred to any location where an e-commerce transaction might take place. Thus, a patent filed 17 years ago when e-commerce did not exist, ended up causing several lawsuits.

The lack of clear boundaries in software means that even law-abiding software developers who intend not to violate another's patent have no clear means of avoiding it. The authors point out that there are around 4000 patents on e-commerce and around 11,000 patents on online shopping. Add to this the fact that getting legal opinion on each software patent can cost around USD 5,000 and we have a vexatious, if not impossible, task at hand. For most software developers, doing a patent search in connection with their work is simply not economically feasible. Even leaving aside the cost of a search, the results are seldom conclusive. Thus it really is not possible to eliminate the risk of a patent infringement lawsuit.

It is well known that the U.S. has the most permissive patent system in the world. However, even in the US, there are signs that the pendulum may be swinging the other way. In the recent Bilski case, which dealt with a method of hedging risks in commodities, the US courts ruled that abstract ideas which are not tethered to a device cannot be patented. The decision reversed the 1998 State Street decision that opened the floodgates for software patents.

In the European union, a move to patent, “computer implemented inventions” was thrown out in 2005. In India, section 3(k) of the Indian Patent Act says that, “A mathematical or business method or a computer programme per se or algorithms are not patentable.” In the discussions around India's Draft Patent Manual, the interpretation of the term, “computer programme per se” has been the most contentious one. Given the lessons of history and considering the amount of litigation that software patents have created in the US, India must avoid going down the same path.

A patent is a state-granted monopoly on an invention, in return for disclosure of the idea. The original intent of the patent system was to encourage disclosure by the inventor in exchange for exclusive rights for a limited period of time to the invention. This ensured that inventors did not take their inventions to the grave and that society could build on existing knowledge rather than re-invent the wheel. The regime of software patents began its major expansion in the 1980s in the US. Since then, software developers have been consistently arguing that that software is better protected through copyrights rather than patents.

Under copyright law, if software developers write code that is similar to that of another, they can defend themselves on the grounds of independent invention because copyright protects the expression of an idea. However, the same defense is not possible under a software patent regime because a patent is a monopoly on the idea itself. Thus, even if software developers independently create a program, they may be liable for infringement of one of the more than 200,000 software patents in existence in the U.S. Even end-users who use software for routine, everyday activities may be liable for infringement. For example, McDonalds and 400 other entities were served notices for violating DataCard's patent on “Method for processing debit purchase transactions using a counter-top terminal system.” In another case, a company (ironically) called Beneficial Innovations, sued the New York Times, You Tube and many other media organizations for allegedly violating its patent on “Method and system for playing games on a network.” Therefore the problem of software patents is not one that is confined to the software development industry alone and ends up increasing the cost of software for society as a whole.

The problem is compounded by the fact that litigation is an expensive process. Dan Ravicher of the Public Patent Foundation points out that for a patent holder to send a cease-and-desist letter, all it takes is a post card. However, that inexpensive post card sets off an expensive chain of events for the defendant who will typically pay a lawyer USD 40,000 to get a legal opinion, around USD 2-4 million in attorney’s fees if the case goes to court and many millions more if the defendant loses the case in court and is required to pay damages.

The argument in favor of software patents is that patents promote innovation. The social contract between an inventor and society was that the inventor disclosed details of the invention in return for the patent, and this disclosure would lead to future inventions. However, the history of the software industry shows that innovation flourished long before software patents came into force during the 1980s. Some of the fundamental inventions of the computer age—the Internet, compilers, spreadsheets etc--were created despite the lack of patent protection. It is therefore clear that patent protection is not necessary for innovation in the software industry.

As with any other monopoly, a patent must be treated with great discretion, especially since this particular monopoly is bestowed by the state itself. The act of granting a 20 year exclusive right to profit from an idea to a private entity needs to be weighed against the cost that it imposes on society. Since software and business method patents prevent independent invention, do not function well as a system of property and lead to increased litigation, India must comprehensively reject it.

Update: Micheal Tiemann, president of the Open Source Initiative and my colleague at Red Hat has some generous praise for this article and why Indians should heed its message.