Monday, May 23, 2005

Software patent debate

This is a lengthy and amazingly well-informed debate about software patents in Europe from Lessig's blog. Seeing who the participants are, this should come as no surprise. It is nice to see some of the best proponents of software patents (I refuse to use the euphemism in vogue) argue their case against some well-informed opponents. One of the main criticisms that I have against the critics of software patents is that often they use over-simplifications of the arguments. I understand that it is difficult to convey the nuances of the EPO's board of appeal rulings to the masses, but using downright fabrications is not the solution.

To me, the case is still an economic one. What is the economic case for software patents? Is there a real need for software patents in Europe? I maintain that the answer is negative.


kc said...

I find it useful to divide the software patent debate into internal criticisms (ie of the patentability of software) and external criticisms of the patent system itself.

Internally, I think the argument against the outright prohibition on software patents is poorly justified. If software embodies a non-obvious useful step such as would normally give rise to patentability but for it's status as software then that must be an arbitrary distinction to make. While I disagree with the logic, I do think the strongest argument against software patents on this front is the numerical/grammatical structure of software.

However, I still do not support software patents, because I think they highlight the problems of the current patent system, and are particularly susceptible to the inefficiencies it generates. The patent system works adequately for discrete inventions, but such inventions are increasingly in the minority and this is particularly so in the technology sector, where an end-product may embody many patents. And so we have seen the rise of cross-licensing and patent portfolios, both of which are products of the system's failure, and which pose a signficant risk to those in the weakest bargaining positions (ie small-scale developers).

The point is this; software patents are objectionable not because software is somehow outside the scope of patentability, but because the quick development cycles and low cost model will expose most easily the disproportionate protection that the patent provides.

Andres Guadamuz said...

Excellent comment! I have spoken with a couple of strong proponents of software patents (American style), and they usually start the discussion by admitting that there are many obvious patents out there, and that this is a problem, but that we have to stick with the system as it is and fix the patent offices.

There are loads of obvious patents because the rules are not clear, and the nature of software development makes it difficult not to have many different people producing the same solutions all the time. The U.S. patent environment has become a race as to who can patent any obvious software tool just by using technospeak and euphemisms.

kc said...

Reform of the Patent Office is superficially attractive, but ultimately it's a question of how much you want to spend and who bears that cost. Strengthening could be achieved by increasing the scrutiny of patents, which costs more, and which would probably raise the cost of applying for a patent. Alternatively you can adopt a very permissive regime that relies on the courts to weed out bad patents, with the cost of litigation born by the parties. As we've seen the costs involved are prohibitive to the extent that it's more sensible to patent as much as possible, and then sit down with competitors and cross-license to avoid litigation.

Another way might be to require the periodic renewal of patents at increasing costs, so that while bad patents may be granted, the economics of maintaining them will ensure that only very strong patents reach the 20 year mark.

Again though, all of these expose the fact that the strengthening the patent system will only help towards those with deep-pockets. A large company, say Microsoft, can afford to maintain and litigate over a large number of strategic patents, without the current problem of small-scale patent holders holding up development.

I think the third option is most attractive, but really, what I'd like to hear is someone propose a workable alternative to the current patent system. And so far I've not see one.