Against Software and Process Patents

I have watched in great sadness as well as some very real fear for my profession as I've seen software - and worse: process - patents gain hold in this country driven by forces of great wealth and power aimed at maintaining their wealth and power at the cost of innovation. I was in the software labs of the late '70s when things like object oriented programming, bitmapped displays, email, and modern operating system theory were being developed and nothing was even copyrighted. There was great excitement as people built on each other's work, creating new concepts - like inter-process communication (IPC as it was known) that are now mainstays of every computer and network in existence (even your cell phone has IPC in it, but it's so taken for granted these days that few even know about the great efforts that went into the creation of shared memory and socket-based communication systems).

In '79 the military got wind of some fantastic new machines - Lisp Machines - that were capable of processing feats way beyond traditional processors. They paid the primary inventors $11 million (if memory serves me correctly) to copyright every piece of software they could that went into those machines. This was born Symbolics, known to some as "Slimebolics," and the great openness of the M.I.T. Artificial Intelligence Laboratory, that until that time had no locks on any doors and no passwords on any computers, began to come to an end as a direct result of the desire to hoard this new resource known as "software."

In the mid-1980's software patents began to appear. I was doing some time at Oracle and was brought into a meeting with the Oracle lawyers who explained to us that if we were working on anything patentable, that we should bring it to their attention as a candidate for a possible software patent. I was outraged and, though Larry wouldn't listen, Oracle's vice president and original architect of the Oracle RDBMS, Bob Miner understood the implications and became an outspoken critic of software patents, even testifying against them to the U.S. Congress. One reason he gave was that Oracle, the second largest software company in the world, would not exist if the IBM RDB code had been patented, as Bob used that as a blueprint for the first Oracle database software. Oracle finally set a policy in place that they would continue to acquire software patents, but not use them against any one who was not itself using them against others. This innovative "no first strike" was overturned by Larry Ellison a few months after Mr. Miner's untimely death.

So now we not only have patents on software and business models, but also on medical processes that can prevent a doctor from helping their patients for fear of being sued for patent violations. Finally, the Supreme Court is starting to take notice; yesterday they heard arguments on a patented blood test that could save lives if freely available. This is a landmark case and could potentially have far reaching consequences. I can only hope that it marks the beginning of a return to some measure of sanity, though there is so much money/power/momentum behind the current broken system that it will at best take years for things to straighten out.

The bottom line is that software patents, as they are currently structured, severely hamper innovation growth in the state of the art (unless you are well funded enough to routinely run $20,000 patent searches on any new ideas that you have, and if something comes up, have a deep enough patent portfolio to enable mutual patent licensing to prevent possible litigation.

Well, that's a little bit about how I feel about the subject. For more, see these articles: