Brian W. Carver -- Share Alike

Arguments Against Software Patents

Brad Templeton writes on his blog, Brad Ideas, that software and hardware are the same thing and so it bothers him that some would allow patents on hardware but disallow them for software. What follows is the comment I posted on his blog.
--
I like your idea for filtering out some of the bad patents.

You're also right that some people simply say, "we should not allow patenting of software." I would like to put forward a couple arguments for that conclusion, rather than simply state it. Then I also have a question for you.

A few arguments against software patents off the top of my head:

0. Since software is already afforded copyright protection, which lasts the life of the author plus seventy years, software production is already adequately incentivized. To additionally provide patent protection to software drastically skews the bargain with the public that all "exclusive rights" bargains seek to achieve and instead provides the software developer with an unnecessary windfall.

1. Software patents have a disparate negative impact on the development of Free software, which for other public policy reasons we would prefer to promote. Free software developers, unlike proprietary software developers, are typically fiscally unable or philosophically unwilling to license patented software techniques. The benefits that Free software provides to both businesses and individuals are worth more to the public than providing a monopoly right to a software patent holder.

2. Software, by its very nature, is susceptible to fewer techniques for achieving its ends than are many other industries. Consequently, where in other industries one can "design around" a patented technique in order to achieve the same end and to thereby compete in the market with the patent holder, "designing around" in the software world is often much more difficult or impossible. There is often simply one or only a few reasonable or efficient ways to accomplish a given software task and if patented, unreasonably enormous amounts of software would be infringing or require licensing.

3. Software cannot be distinguished from mathematical algorithms, and we have traditionally not allowed the patenting of mathematical algorithms because we recognize how stifling to innovation such a practice would be.

4. The software and computer industries advance at a pace totally unlike other industries where we offer patents. Consequently, providing a 20-year patent term for software has a totally different impact in its industry than such a term does in other industries. This is not an argument for the abolition of software patents, but instead suggests that if we are to have such patents, they should perhaps only last two to five years. (How much software do you have from 1984 that really needed to be protected by patents these last twenty years to serve as an adequate incentive for its authors to produce it? None.)

Question: This isn't the first time I've heard a knowledgeable person disclaim any distinction between hardware and software. But, just as you're dissatisfied with those who provide no arguments for not allowing software patents, I can't accept this lack of distinction without a persuasive argument. Do you have one? It seems to me that there are several principled distinctions between the two.
--
(I should have also noted that another way to avoid the inconsistency that troubles him is to abolish patents altogether. This would probably be a better course than allowing software patents, so far as I can tell.) For rhetorical effect it would probably also have been a good idea to include Bill Gates' famous quotation:

"If people had understood how patents would be granted when most of today's ideas were invented and had taken out patents, the industry would be at a complete standstill today."

We're headed that way now.

Posted by Brian at April 9, 2004 07:34 AM | TrackBack