From Software Patents to Silence

A Thought Experiment

Could this be the beginning of the end of software patents? Let's hope so!

Why would more and more people—especially young people—abandon music on the radio, on TV, and on the internet. And why would those who wanted to listen to live music from new and local bands have to do so in secret, and in fear of prosecution?

The U.S. Supreme Court is currently dominated 5-4 by conservative justices. The dominant group have already demonstrated an eagerness to step beyond the bounds of legal interpretation and to seek to impose a radical conservative agenda. To take just one recent example, the Court has ruled that corporations can spend unlimited amounts in U.S. elections (and that includes foreign corporations), thereby overturning more than a century of precedent and practice. (See The "Devastating" Decision.)

In the current legal climate it is quite possible that the Supreme Court may choose to fully validate software patents in law. (See Sawyer on "Why Bilski Really Means That Software Companies Should Leave the US".) This short essay is concerned with what such a ruling could lead to in practice.

The USPTO (U.S. Patents and Trademarks Office) is already granting thousands of software patents annually, despite their questionable status in law, so with a Supreme Court decision this number would be expected to increase dramatically.

Most software in use today is built using a vast amount of pre-patenting prior art. This means that new software can be created using unpatented techniques and this would continue to be possible. Unfortunately, it is practically impossible for individual programmers, in-house software teams, SMEs and other organizations that don't have considerable legal expertise and financial resources to be able to check that any new code they write does not infringe any of the patents that have already been granted—or that may be granted in the future! (Even if you have created the prior art yourself you need significant legal and financial muscle to prove it.)

With software patents firmly established in U.S. law it will only take a few legal clashes to make software development uneconomic for all SMEs and individuals. Software entrepreneurship would no longer be possible in the U.S., at least not outside a big protective corporation. This would lead to a stranglehold on software development by a relatively small number of giant software companies. Entry into the software development market—or even in-house software development by non-software companies—would become prohibitively expensive because of the legal checks that would be required and the legal battles that would have to be funded or at least insured against. Some large non-software companies would pay to play—companies like GM—but anyone smaller would have to pull out.

The software giants themselves would probably do zero-cost cross-licensing deals with each other. This would give them the freedom to innovate, while ensuring that no one else could enter their marketplace.

The Free Software Foundation and similar bodies might try to buy indemnification insurance but their economic position compared to the software giants is so puny that all small-scale and individual software development outside the software giants would cease within the U.S. within a year or two at most.

The European, German, and UK Patent Offices are already idealogically pro-software patents, and are either granting or at least supporting software patents despite their legal questionability. Once software patents become fully valid in U.S. law, the European, German, and UK Patent Offices will be able to finesse pre-existing WIPO agreements to bring them into force in the EU. The net effect is likely to be that the U.S. software giants dominate the EU software market, perhaps with one or two EU-based giants joining in.

Other countries that have or are likely to have strong software bases (right now, Brazil, China, India, and Russia), may well take a two-pronged approach to the issue. They may do bilateral agreements such that they participate in the patenting regime for all products they export to the U.S. and EU, while taking a looser approach for their domestic markets to foster innovation and entrepreneurship. This will most likely swing the centre of the software development universe out of the U.S.

Few software developers seem to be aware of the danger and people outside the software industry mostly don't know and don't care. But one thing that people do care about is music.

Any piece of music can be represented by a computer program. The program might actually play the music, e.g., using the computer's "bell" or using a sound card. Or it might hold a representation of the written form of the music. And the reverse is also true: any program can be represented by a piece of music by performing a suitable mechanistic translation from code to music (as music notation or as a sequence of sounds). This means that once software patents are established, so too is the principle that music itself is patentable.

Music has far more prior art than software. Nonetheless, we could expect to see the same pattern emerge. The big music labels would cross-license each other and squeeze everyone else out, just like the software giants described above. No independent bands or individual artists could afford the legal fees to check or defend their compositions, so no one in the U.S. and EU could make music as a free agent.

Links: lots of information on software patents; a short informative video on software patents; Europeans against software patents; the Electronic Frontier Foundation on software patents; Russel Winder's blog with lots to say about software innovation. And it seems I'm not the only one to see software as a form of mathematics: 1+1 (pat. pending) — Mathematics, Software and Free Speech. See also, Appendix B: The Dangers of Software Patents, in my book Programming in Go.