The Congress shall have power: ... To promote the progress of science and useful arts by securing for limited times to authors and inventors the exclusive rights to their respective writings and discoveries.

(US Constitution, Article I, Section 8)

I'm all for promoting the progress of science and useful arts, but I'm not convinced the current practice of patenting software moves us very far in that direction. Quite the opposite, in fact. Software patents, if they stand up to legal scrutiny, could well hobble the software industry as a whole, stifle innovation in commercial software packages, and cripple free software irretrievably.

Patent Economics

Large companies generally have few problems with patents. They accumulate large pools of patents, both for possible revenue and for self-protection. If threatened by another company, they hit back with claims of their own. In the end, they generally end up cross-licensing everything.

There was an amusing example of this sort of exchange a while back. Apple was suing IBM for some sort of infringement, and IBM retaliated with their favorite hole card: DMA. That's right; IBM owns the rights to Direct Memory Access, used by essentially every computer system made in recent history. Unfortunately for IBM, Apple hadn't installed DMA hardware in its boxes. Instead, all the I/O was run through the CPU chip: slow, but real cheap. So, one hole card blown to smithereens; try again, IBM...

The story gets less amusing when one of the players is a small company. If Apple decided to sue my company, there is very little I could do to defend myself. I have no pool of patents, let alone cross-licensing agreements, and lawyers are notoriously expensive. The result, effectively, is that my comapny (and many others) are locked out of the patent-holding cartel.

Smaller companies have no effective way to check for patent infringement, and no real chance of defending themselves in court. We could easily see the software industry becoming ever more monopolistic, simply because of the current patent situation.

Patents & Freeware

This is hard luck for small companies (and their customers), but it could be devastating to the future of free software. Freeware authors have no revenue streams to bank against possible lawsuits. Can they take the chance of releasing something that might cause a lawsuit? Should they go underground, releasing software only through anonymous remailers? This may seem a bit far-fetched, but the dangers are quite real.

Freeware redistributors are much more exposed, due to their fixed addresses and the vast amounts of material their collections can contain. Five gigabytes of source code leaves a lot of room for infringements, and there is no reasonable way to check out that much material. For that matter, FTP archives aren't immune from lawsuits, and they make no money at all from their services.

How Bad is It?

Pretty Bad. The Internet Patent News Service patents@world.std.com has finished its analysis of US software patents issued in 1994. It found a final total of 4569 patents, pushing the total software patent count since 1970 over 15,000. The five hottest areas in 1994 were Image processing (623), Networks & communications (532), Operating systems (448), Process & numerical control (374), and Graphics (337). Other familiar areas include Graphical user interface (223), Signal processing (211), Database (173), Security & encryption (162), and Word processing (121).

The top dozen companies, by the way, were IBM, Hitachi, DEC, Xerox & Fuji Xerox, Toshiba, Hewlett-Packard, Fujitsu, Canon Kabushiki Kaisha, Motorola, Matsushita Electric Industrial, Mitsubishi Denki Kabushiki Kaisha, and ATT Bell Laboratories. The top dozen companies accounted for over a quarter of the pateents issued, and Bell Labs (remember Bell Labs?) is in twelfth place! The list is clearly dominated by BIG American and Japanese companies; this game isn't being played to protect garage-based inventors!

Nor, I'm afraid, is it concerned with promoting "the progress of science and useful arts". The software patent explosion is really a colossal land grab, and the government inspectors aren't watching the claims all that closely. As Gregory Aharonian, Editor of the Internet Patent News Service, says:

Analysis of the 1994 software patents in light of the largest software prior art database in the country (maintained by Source Translation & Optimization, publisher of the Internet Patent News Service) shows that over 75% of the 1994 software patents should have had one or more of their claims disallowed on prior art grounds.

Even if only 25% of the patents got through, however, we would still have more than 1000 new patents to worry about. Programmers can't reasonably be expected to do patent searches every time they write a new piece of code. Worse, because of the delays in the patent system, a piece of code could well be on the market for years before the patent was granted (and made public).

What Can We Do?

Not much, as individuals. The giant corporations are having a field day, and the government seems to be staying out of the way. Aside from keeping the problem alive in discussions, objecting to company efforts to patent trivial and obvious work, contributing to archives of prior art, and writing the occasional letter, individuals are pretty much overpowered in this battle.

Collectively, we may have a bit more success. Members of organizations such as ACM and IEEE can push to have statements made about the dangers of software patents. We can also support the League for Programming Freedom (1 Kendall Square, #143, P.O. Box 9171, Cambridge, MA 02139 USA; lpf@uunet.uu.net; http://www.lpf.org). The League has been fighting against software patents for years; now is a great time for more programmers to join their ranks!