Devil's Advocate: That is my software!

Is the creation of software innovative enough to require legal protection? There's certainly a right to reasonable return but, says Martin Brampton, current copyright and patent laws aren't the answer

SCO is claiming that it owns a big chunk of Linux. The trouble is we don’t really have any clear understanding of how to manage rights to software. We don’t even know whether there should be any rights at all.

Two mechanisms have been the main tools for people wanting to control software. Neither looks very suitable. Copyright was the first to be extensively applied, although its use with literary and musical works is far easier to understand. Recently, attempts to patent software ideas have been popular.

Copyright is a modern innovation, with the first directly relevant Act of Parliament enacted in 1709. Its motivation does have some resonance with modern circumstances. Pressure for legislation came from the huge decrease in the cost of producing books brought about by printing. Piracy became rife when copying no longer involved laborious writing by hand. Today, with the reproduction and publishing of electronic texts involving practically no effort, one might suppose that copyright is even more important.

But copyright for software is looking problematic. When people first started writing copyright marks on software, programs typically contained a few dozen modules, each with no more than a handful of pages. Now it seems that every piece of software has millions of instructions. That makes the application of copyright difficult.

In copyright disputes over texts or music, it is usually quite easy to understand what the argument is about. Often the issues can be made plain to the general public by offering illustrations of the original and the allegedly pirated work. This is not so with modern software. SCO claims that a million lines of code have been lifted from its copyrighted material. Yet I have heard of nobody who has the slightest idea of what code is involved, or is able to form an informed judgement on the merits of the claim.

Patents for software raise at least as many difficulties as copyright. Much basic material has long been in the public domain. Many commentators say that patents are being granted far too readily, simply because nobody is in a position to challenge them. It is hardly practical to scrutinise a piece of code and then to compare it with everything else that has ever been written to see whether it is genuinely innovative.

Protection by patent was originally applied mainly to innovative pieces of machinery. Part of the justification was that it took significant time and investment to start to manufacture something new. Without protection, the effort would not be worthwhile. This is not really true of fragments of program code.

That brings us to the fundamental issue of principle. Is the creation of software innovative in a way that requires legal protection? Certainly the writing of software is a creative act. But it also depends very heavily on past creativity. A programmer who understands a good selection of efficient algorithms is likely to be more innovative than one who knows only the programming language.

Trying to isolate individual acts of creativity therefore seems invidious and counter productive. All of us depend very much on everything that went before in order to add our own modicum of creation. Moreover, much programming is routine, with only some elements requiring special effort. Even then, one suspects that there are a good many skilled developers who could, given the same problem, come up with good solutions quite quickly.

We could argue that the creator of software should have a right to a reasonable return. But that could cut two ways. If there is a right to a reasonable return, why is there not a right to limit the return to a reasonable level? That only emphasises the fact that notions of ownership of intellectual creations are purely man made. Might we not be better off simply refusing any protection to software and hoping to benefit from each innovation being applied as widely as possible? That would certainly stop the SCO argument.