As readers of this blog will know, software cannot be patented in Europe “as such”; quite what that wretched “as such” means is the subject of major arguments. As I noted earlier this week, the European Patent Office is currently conducting a consultation into that and much else concerning the patentability of software, to which it is relatively well disposed.
The UK patent office, which now rejoices in the impossible title of “Intellectual Property Office” - impossible because patents are intellectual monopolies, not intellectual property – has, to its credit, been more sceptical about what might be patentable in the field of software. Alas, the intellectual monopoly maximalists have been pushing case after case through the UK courts in an attempt to wear down that resistance, and it looks like they're winning. For example, I wrote last year about the Symbian judgment, and that has now fed through into what amounts to a fully-blown software patent being granted here in the UK.
It concerns a decision regarding an application from Nokia entitled “A method of rapid software application development for a wireless mobile device”, which was filed on 8 November 2004. As the decision explains:
During the course of examination, the examiner issued several examination reports and raised a variety of objections including novelty, inventive-step, added matter, support, clarity, and excluded matter objections. All those objections were overcome to the examiner's satisfaction except for the excluded matter objection, namely that the invention is excluded from patent protection because it relates to a method for performing a mental act and/or a program for a computer, contrary to section 1(2)(c).
In other words, this goes to the very heart of software patentability.
Here's how the hearing officer describes the purported invention:
a software implemented method for developing networked applications for a wireless mobile device, the software enabling a developer to use a computer remote from a wireless mobile device to call, over a network connection, modular software elements resident on the wireless mobile device and to combine and execute modular software elements resident on the device by using a script composed on the computer and transferred to the wireless mobile device.
Hardly sounds very innovative, does it? Anyway, the nub of the problem is whether this represents a “technical contribution”. This concept is used by intellectual monopolists to turn unpatentable software into patentable software: the difference being that, roughly speaking, “patentable” software makes a “technical contribution” - that is, *does* something that has an effect on the outside world. The question, then, is whether Nokia's super-duper method actually did anything.
Here's what the examiner had to say:
The Court's judgment in Symbian itself also gives further guidance on what constitutes a technical contribution. In finding that the Symbian invention did indeed make a technical contribution the court said “... not only will a computer containing the instructions in question “be a better computer”, as in Gale, but, unlike in that case, it can also be said that the instructions “solve a “technical” problem lying within the computer itself”. Indeed, the effect of the instant alleged invention is not merely within the computer programmed with the relevant instructions” (paragraph 54, emphasis added)"
because
“The effect of the alleged invention in the present case improves the speed and the reliability of the functioning of the computer” (paragraph 55) and “there is more than just a “better program”, there is a faster and more reliable computer” (paragraph 56).
So, the whole issue of patentability hinges on whether an alleged invention is just a “better program” - not patentable - or produces “a faster and more reliable computer,” which is patentable according to the latest judicial tests.
But what exactly does “better program” mean? There's certainly no general metric for measuring whether programs are better or worse than others. Instead, they are judged on whether they *do* things better – you know, things like making it relatively faster or more reliable. Oh, but hang on, that's supposed to be the crucial difference between patentable and unpatentable software.
So, it turns out, that in real life – not the abstract world of court rooms - there is *no* discernible difference between patentable software and the unpatentable kind. The examiner, bless his cotton socks, may have attempted to perceive one, but if he'd asked a few programmers he would have found that he was attempting to create a false distinction – just like all the others based around the fictitious concept of “technical contribution”, and the meaningless “as such” phrase.
Once again, the UK patent office has found itself trying to find a neat, sensible way of testing whether software is patentable or not – and failing, for the reasons I've given above. There is no way of telling because there is only one kind of software – the kind that cannot be patented because it's just a bunch of algorithms – maths – that can be used to make programs produce results, now slower, now quicker, depending on how you use those algorithms.
You either allow *everything* to be patented – and that way lies madness and astronomical legal bills as the US has shown – or you allow nothing. I know which I'd choose...
Follow me @glynmoody on Twitter or identi.ca.