Here's the notes I (Malcolm Tyrrell) used to argue the case against software patents. I was the second speaker on this side of the argument and took an economics perspective. The first speaker on this side had discussed isses from a programming perspective. Note that the actual speach I gave didn't follow the notes below precisely.
Our side won the debate.
Most people are well aware that mechanical inventions, pharmaceuticals and various other technological creations can be patented.
In fact, most people are suprised to discover that software is not similarly patentable.
However, software is radically different from other fields of technology and for this reason a one-size-fits-all justification for patenting software isn't sufficient.
So why is software different?
I'll identify three main differences:
Firstly, a piece of software, unlike a machine or a drug, does not depend on any laws of physics, or the biology of the human body.
The author of a piece of softare does not have to worry about the imperfections of the everyday world.
Instead, the world of software is a mathematical virtual world.
In this sense, software has much more in common with mathematics than with other fields of technology.
Secondly, software gets legal protection under copyright law, yet the end products of other fields of technology do not.
This in particular suggests a significant difference of kind between software and other fields of technology.
The third difference, and the most relevant for my argument, it is comparibly easy to develop.
To bring a physical device to market requires a massive investment in manufacturing infrastructure.
To bring a drug to market requires expensive certification trials, that sometimes fail to validate the drug.
To write software requires a computer and some people.
Given these radical differences, the question of whether software patents are a good thing or not needs to be considered on its own terms.
What purpose do patents serve?
The main role of patents, nowadays, is to act as an incentive to innovation.
This incentive is in the form of a temporary monopoly on the design of a product, or drug or whatever.
The government offers this incentive and in doing so artificially intervenes in the marketplace.
You don't need to be a free-marketeer to appreciate that this has several negative consequences.
Monopolies are, by their nature, uncompetitive.
New expenses are introduced: legal expenditure.
New risks are introduced: companies can be sued for patent infringements.
Furthermore, in granting the patent, the government trades away a tiny amount of the freedom of its citizens: the freedom to build and use comparible products.
However, in some markets, this intenvention can arguably be justifiable:
Let's briefly consider pharmaceuticals:
The design, testing and susequent certification trials of a drug is very expensive.
A company which invests money in the process of bringing a drug to market, needs to know that its competitors cannot directly utilize this development.
If its competitors could, then the company would be at a huge disadvantage: having spent money on developing the drug, it has less money to invest in marketing it.
So there would be a natural unwillingness develop pharmaceuticals without the government intervening.
All this makes, at least as far as I'm concerned, a reasonable case for having patents for pharmaceuticals.
However, the case for goverment intervention in the software market; the creation of these artificial monopolies, and the consequent expenses and risks; is different, because software is different.
We should accept software patents only if the negatives consequences for software innovation are outweighed by the benefits.
The main reason why this is not case, is that to write useful software is not very, very expensive. It is relatively speaking, cheap.
Because of this, far less incentive is needed for a company innovate in the software market.
The negative effect of the expenses, the risks and the monopolies on software dwaft the positive effect of patents as an incentive.
Compared to total investment in the development of a drug, hiring patent lawyers, getting them to search for existing patents, translating ideas into comedy patent language, and the cost of filing patents are small.
Compared to the total investment in a software project, these costs are high.
Given the total investment in the development of a drug, the costs of searching for and licensing existing patents are small.
Compared to the total investment in a software project, these costs are high. If you're lucky enough to discover the patents that you might infringe.
Given the total investment in the development of a drug, the risk and consequences of being sued for a patent infringment are, relavitely manageable.
Compared to the total investment in a software project, these costs are prohibitive.
These arguments hold true regardless of the types of software in question:
Desktop software, the software on a web server, the software which controls an ABS system in a car, the software which mananges communication protocols in a mobile phone, and so on.
By exposing most companies to new and unnecessary expenses and risks, software patents will make innovation in the software market more costly.
Of course, not all companies will be losers.
Very big companies, for example, can benefit and use patents to shore up their existing monopolies. I don't have time to persue this point, but I can point out that in large corporate court cases, the company with the deeper pockets has a disproportionate advantage.
I do want to mention a new and disturbing development that's taking place in markets where software can be patented.
Recently, a new type of company has come onto the scene.
Its sole asset is its patent portfolio and its only employees are lawyers.
Its business model is to buy patents, typically software or business process patents, off dying companies on the cheap and then use them to sue other companies.
This type of company does not develop software.
It is entirely parasitic on the patent system.
The existance of such a company is an obvious indicator that there are problems with software patents.
Overall, software patents will stifle innovation.
Since the purpose of software patents is to act an incentive to innovation, they won't work.
In closing, where the EU to allow software to become patentable, software developers in Europe will be exposed to unnecessary expenses and associated risks.
To be honest, I'd rather invest my money in writing better software.