Patent nonsense revisited

I really can't believe - can barely understand, given its dense legalities - this latest ruling, which seems to want to reinstate software patents in the UK. I had thought that the reasons software should not be patented, but given instead other sorts of intellectual property protection, were by now widely understood.

Perhaps Justice Kitchin has been talking to a friend of mine with whom, years ago, I had a knock-down argument that neither of us won.

My friend's engineering nous is exemplary, his creativity is off-planet, his logical prowess would shame a Vulcan, and his commercial instinct is so minuscule I doubt CERN could detect it with its finest machinery set to maximum magnification.

In his work for an organisation you will encounter every day, my friend built an impressive portfolio of patents in some very technical fields. (I think one particular set is about to lapse, just at the moment that the ideas behind them are about to become very important. If it were anyone else I'd find that immensely sad, but with him it's as natural as apples falling from trees.)

He was of the opinion that of course, software should be patented: he had created hardware-based inventions and software-based inventions, and couldn't see why one set was protected and the other not.

I completely understand that, from his point of view. He had created complex things that had never been seen before, doing things that nothing else did.

But pure software is so rarely like that - there are immense commonalities between techniques, because at heart everything is based on the same tiny set of mechanisms, and the problems solved boil down in a remarkable fashion to different ways of expressing the same problem and applying similar solutions.

I remember when I was involved in the initial design of an Office-like suite of software about twenty years ago, when the field was a lot newer than it is now and there was more scope for philosophical discussions of what it was that such a task actually involved (for those with extreme nerdular tendencies, there are tenuous connections between those discussions and what ended up as PipeDream on the Z88).

We decided that everything - word processor, graphics, spreadsheet - could be considered as special cases of a database which stored just two things: information contained in a small variety of data types, and interrelationships. The word processor would work with type 'character', building up into relationships called words, sentences, paragraphs and so on, with those relationships stored as their own types with relationships covering formatting, document intent, and so on. Spreadsheets would know about type 'number' interrelated by formulae and formatting, and so on and so forth. The freedom for mixing up what these days are considered very different sorts of things was startling, and had such a thing worked well we'd probably have a very different online world now.

As you might expect, philosophically satisfying schema like those soon hit some ugly realities about performance, storage and other dull old practicalities. (Whether it's worth revisiting these ideas now we have stupidly capable machines is in itself a fascinating idea). But I think it would have been possible, had software patents been in existence back then, to have written a single patent that would have knobbled enormous areas of application software for years, without any recourse to actually producing an invention that had much practical application.

My friend refused to believe that such worries had any place in real invention, because he really believed that patents existed purely to encourage innovation and would be used as such by anyone who was a real inventor.

I so wish he had been right. I was unable to convince him otherwise. But he wasn't, as the existence of patent trolls demonstrates, and until we can legislate against bad faith we'll have to content ourselves with legislating against software patents.