Print

Stallman on the Harm of Patents

An Exchange Between Richard Stallman and Tim O'Reilly

03/13/2000

Tim O'Reilly discusses software patent issues with Richard Stallman, founder of the GNU Project and Free Software Foundation. This dialogue took place as an exchange of email from February 29 through March 2, 2000.

Tim O'Reilly: I'm not completely opposed to software patents, since there are some things that do in fact qualify as legitimate "inventions".

Richard Stallman: I hope you will reconsider this step in your reasoning. Yes, there are real inventions in software. I am impressed by the RSA algorithm, and even the LZW algorithm, as intellectual work; they are far from obvious, in my opinion. But look at how much harm these patents have done!

O'Reilly: I agree completely, and that is the starting point of my reasoning. Here's how I see it:

1.) Leaving aside the business effects of patents for a moment, let's just consider "good" and "bad" in the context of invention. Clearly, a patent like the RSA patent embodies serious, original work in computer science that qualifies under the traditional understanding of a patentable invention. Equally as clearly, something like Amazon's one-click or associates patent represents a completely trivial application of prior art. The simple test: ask someone to create a public key cryptography system, even today when the technology has been around for many years, and you'd be hard pressed to find all that large a group of people who could re-implement something equivalent to the RSA work without reference to information about their work. By contrast, ask any web developer to implement something like One-click or an Associates program, and there are hundreds of thousands of people who could do it in an hour, without any reference to anything done by Amazon. What's more, this was almost as true when Amazon filed their patent (though the numbers of developers was smaller back then, because the web itself was smaller.)

Stallman: I basically agree with you this far, on the issue of judging new developments as to how much intellectual admiration they deserve. But when we consider public policy on software patents, we should not treat them as medals to be handed out to inventors who deserve admiration. Much more is at stake.

O'Reilly: 2.) What's more, these trivial patents undermine the system in a far more insidious way. Precisely because they are so trivial, they effectively amount to a patent on an idea, rather than on an implementation. And regardless of whether patents in general are good or bad--a point on which reasonable people can differ--it's clear to almost anyone involved in the computer field (except the robber barons who are trying to take advantage of the weakness of the PTO in dealing with new technologies) that broad, business-process patents, which amount to the patenting of fundamental ideas rather than of any real invention, is bad.

3.) You're absolutely right that even credible patents, like the RSA patent, can do great harm. In his book Owning the Future, Seth Shulman points out that the Wright brothers' early patents on aviation held back the development of the industry in this country versus that in Europe, where their patents didn't apply, to such a level that the government had to step in and nationalize the patents. However, it is clear (at least to someone with only a generalist's knowledge of the subject) that in some areas, patents do spur innovation. There are certain expensive fields of research that are undertaken only with the hope that there will be some protection for the result.

Stallman: There is little hard evidence for this -- but I would tend to agree that in some fields, such as development of medicine, patents are beneficial because the expense of any new development (even a fairly simple system in terms of the number of parts) is very great.

Software is at the opposite extreme, though.

O'Reilly: So I'm not ready to say the patent system as a whole should be overthrown. I am ready to say that anyone who files a patent should think long and hard about the overall business effects of such a patent. In most cases, I'd bet that being restrictive about a patent holds back the market, and so you end up eating more of a smaller pie. That being said, I have always supported the right of true creators of value and ideas to make their own decisions about the best way to profit from their ideas.

Stallman: In so doing, you adopt the view that was rejected in writing the US Constitution. The US Patent system exists, in principle, to promote progress, not because inventors are considered to have a special entitlement to a monopoly.

The paper "Against Software Patents" says more about this, and about why even nontrivial software patents, which are by no means mistakes, are still bad policy.

O'Reilly: In short, I really do want to focus my energy on frivolous patents, a subject on which I am completely clear, and only secondarily on the bad effects of all patents, which, as I've noted, is a subject where there is room for argument. Let those more knowledgeable than I about those arguments carry them out in public, and let's see some thoughtful public policy on the subject.

Stallman: Perhaps you are starting from the idea that "Where there is a real invention, there should be a patent." But why assume that? People who note the existence of absurd patents often assume (without any grounds they can cite) that these are the whole extent of the problem--that the patent system would be ok for software, if only it were competently run. But there's no basis for that belief, except a reluctance to doubt the system.

Under the Constitution, the purpose of the patent system is to promote progress. In the field of software, patents do more harm than good, even the ones that cover real inventions. Have you read the LPF's position paper, Against Software Patents? If not, I urge you to take a look. You can find it easily through http://lpf.ai.mit.edu.

O'Reilly: No, I hadn't. Thanks for the pointer. I assume you're talking about http://lpf.ai.mit.edu/Patents/against-software-patents.html. I was particularly struck by the comments:

"Sometimes it is possible to patent a technique that is not new precisely because it is obvious--so obvious that no one would have published a paper about it."
and
"Patent applications are written by lawyers for lawyers."

This is a real problem for those of us on the technical end. There's a discontinuity between the common sense of the programming community, and the common sense of the legal community. We don't understand the language that they speak, just as they don't understand ours. But I do think that in an odd way, abusers like Amazon and Priceline are doing us all a favor with their over the top patents, since they are making the system such a clear burden on the industry that (I hope) something will have to be done about it.

Pages: 1, 2

Next Pagearrow