Friday, June 22, 2012

EFF right idea but wrong target on software patents

The EFF has started a campaign against software patents. They state the problem as:
"Software patents are hard to understand. Really hard, because patent lawyers can get away with writing them in extremely vague and broad language. This means that innovators are left without clear guidance on what patents they might be infringing. It also allows trolls to exploit the vague and confusing patent language to extort higher licensing fees than they would otherwise be entitled."
Much though I like the good folk at the EFF I think maybe they've got their target slightly wrong here.  Software patents can be vague and broad but I'm not sure those particular features distinguish them from any other types of patents.  Yes patent holders exploit the language but companies like Intellectual Ventures or Blackboard, IBM or Myriad Genetics don't really care if it is hardware, software or even, these days, bioware or business methods at the heart of a case.  What they exploit, primarily, is the cost of patent litigation system.  Going to court is so expensive it is really only feasible for those with deep pockets.  For smaller companies the patent system is a licence to pay a lawyer.  If one of the bigger corporations decide to target you the only rational economic response is to settle. Otherwise the risk of bankruptcy is severe.

Greg Aharonian has been a vocal and long time critic of those opposed to software patents. He makes the perfectly logical argument that:
"The Church-Turing thesis (for every algorithm, there is a circuit) requires software and hardware patentability to be equivalent, especially in light of hardware-software codesign tools, where you design both (and their combinations) with equal effort."
And he's right.  If our societal values and patent systems legitimise patents of certain types of electronic hardware then there is no logical reason why the software equivalent of that circuit should not be patentable.

My problem with these types of patents, software and  hardware equivalents, is that they amount to patenting algorithms.  Software is mathematics and mathematics should not be patentable.  Now I know this is a much more complicated case to articulate in practice to technically and mathematically challenged policymakers. But the trouble with focusing on simple but erroneous targets is that they lead to simple but erroneous "solutions" that don't fix anything.

The patent system is complicated.  The technologies that it is increasingly covering are also incredibly complicated.  The principle that mathematics should not be patentable is clear but gets lost in translation when the messy dysfunctionality of the patent system gets applied, in real life, to the complexity of modern technology. Too much of the public debate on intellectual property consists of IP maximalists and IP minimalists shouting PR digestible slogans at each other.  My concern is that the EFF's campaign here might just end up in the same unproductive arena, focusing attention on a beguiling but deceptive underlying issue of vague patent language.  Vague and broad patent language is a problem but it is not one that is exclusive to software patents.

No comments: