From: www.itworld.com

An interview with Richard Stallman

by J.S. Kelly

March 26, 2001 —

 

Copyright 2000 LinuxWorld

Verbatim copying and redistribution of this article are
permitted in any medium, provided this notice is preserved.



In December, Richard M. Stallman called for a boycott against Amazon.com due to
its aggressive use of patents against competitor Barnes and Noble.
Stallman, father of the GNU Project, free software activist,
and legendary programmer, graciously agreed to speak with LinuxWorld on how software patents have been a
problem for programmers for nearly 20 years, and how the problem is now
being thrust into the forefront yet again with the recent Priceline and
Amazon.com legal actions. He also talked about patent pools and the League for
Programming Freedom, and about money-squeezers and possible solutions for
the problem of software patents.


LinuxWorld: To start, could I ask you to lay out the basic problems with
software patents?


Richard M. Stallman: Software patents monopolize an algorithm, or a
feature, or a technique so that nobody [but the patent holder] can use
them in developing a program. And this makes software development
dangerous. When you are writing a large program and you're using many
techniques, implementing many features, the likelihood is that some of
them are patented by somebody. Or even a combination of them could be
patented.

LinuxWorld.com links

LinuxWorld.com home

Best of LinuxWorld.com

The Legacy Files

The Penguin Brief

Version Control

Linux links

Linux forums


LinuxWorld: And so what can software developers do to protect themselves?


Richard M. Stallman: They can't. There's just nothing [one] can do.


Large companies can [indirectly] protect themselves, [in part]. What they [can] do
is get a lot of patents themselves; then, if a large company like IBM
is threatened with a patent by some company that really makes things, IBM
probably has a patent covering something that company makes, probably has
many patents covering them. So IBM will make them cross-license.


LinuxWorld: It's been suggested that free software programmers could create a
patent pool to do exactly that kind of cross-licensing for free software.


Richard M. Stallman: It doesn't have to be just free software developers; other people who
want to protect themselves could join such a pool as well. The problem is
in getting it started, because the bigger the pool is, the more beneficial
it is to join. So the smaller it is, the less reason there is to join.
Nobody has ever been able to get one started.


LinuxWorld: And why do you think that is?


Richard M. Stallman: Persuading companies who have patents to join this pool is apparently
hard because nobody has ever done it.


LinuxWorld: Has a patent pool ever gotten off the ground, as a thing that people
can join?


Richard M. Stallman: Nobody has ever formally started a patent pool. I guess nobody has
seen it was useful to go through that work without having people who were
going to join it.


Other problems are that it costs a lot of money to get a patent -- many
thousands of dollars, sometimes tens of thousands. So free software
developers can think about the idea [of obtaining many patents for
contribution to a pool or for cross-licensing], but to actually do it is
not easy.


LinuxWorld: It seems there's a growing sentiment against software patents, what
with the Amazon controversy and all that.


Richard M. Stallman: Well, I hope that this will lead to legislation that will more or less
solve the problem. The best way to solve the problem would be to pass a
law that would put software out of bounds for patent infringement
lawsuits. A similar law was passed to cover surgery a few years ago, when
there was a controversy about patents on surgical techniques.


For many years, there hadn't been any patents on surgical techniques. Then
some people started applying for such patents, claiming that such patents
were necessary to encourage innovations in surgery -- which is obviously
false, because surgeons have been eager to make innovations in surgery
without patents.


But any time people start introducing patents into an area, they ask
people to suppose that no progress will take place without patents. And
the power of right-wing ideology is so great that they will get many
people to believe those suppositions even in the absence of any evidence
for them.


LinuxWorld: Why do you call it "right-wing" ideology?


Richard M. Stallman: The idea that creating new kinds of "property rights" invariably
makes society better is a classic right-wing position. Many conservatives
have blind faith in it.


In any case, the controversy over surgical patents was so great that
eventually Congress passed a law saying that surgeons could not be sued
for patent infringement for their surgery.


So these patents can still be issued, but they don't cause a problem for
surgery. It's a very neat solution. And the exact same solution could be
used for software.


One of the advantages of [this solution] is that it solves the problem of
existing patents as well as the problem of possible future patents.


It also avoids a very tricky problem, which is trying to classify patents
into "software" or "not-software." That is a very difficult thing to do,
because often the same patent might cover some software activities and
some other nonsoftware implementations as well.


So solutions based on trying to distinguish software patents from others
are very difficult. But a law that would exempt, say, "software running on
general-purpose computer hardware" from being the cause of a
patent-infringement lawsuit would avoid the need to classify patents.


LinuxWorld: Do you think it would take actual legislation?


Richard M. Stallman: It certainly would, yes.


LinuxWorld: A court decision might not be enough?


Richard M. Stallman: No, it would require legislation by Congress. That's how a law is
passed, after all. That's how the law exempting surgery from patent
lawsuits was passed.


LinuxWorld: You say that these patents on surgical technique can still be issued,
but cannot be used to sue. Are people still applying for them?


Richard M. Stallman: I don't know.


LinuxWorld: Software patents were one of the things that the League for
Programming Freedom was working on. What is going on with the LPF? It
isn't very active today, is it?


Richard M. Stallman: That's right. The league is mostly inactive, but there is somebody
updating the Website from time to time.


LinuxWorld: And why is it mostly inactive today?


Richard M. Stallman: It ran out of activists to be the leaders. That was the problem. We
never got enough members that we could get in enough money that we could
easily keep this thing going. And at the time, we didn't see a way to make
any progress on the issue of software patents.


Programmers were generally against [software patents], at least according
to the few surveys that had been done, but we just didn't see a way of
putting it on the political agenda, and that was discouraging. Well, now,
as a result partly of Amazon, it's getting onto the political agenda. So
it's on the map.


LinuxWorld: I would think that with all the excitement surrounding free software
these days that there would be more activism in this area.


Richard M. Stallman: Well, maybe now. It was several years ago that the LPF became
basically inactive. We didn't have anyone to be the leader. People didn't
want me to be the leader, and I agreed, because I'm so associated with
free software, and the LPF is not a free software organization. The LPF
was against the monopolies that obstruct software development, not just
free software development, but proprietary software development also. And
so we all concluded that it wouldn't be good for me to be the leader,
because that would tie it too much to free software.


LinuxWorld: Can you think of someone who would be right in that position?


Richard M. Stallman: I don't know.


LinuxWorld: How was the leader chosen back when the LPF was more active?


Richard M. Stallman: The officers were elected by the members.


LinuxWorld: Now that they have lots of capital, a lot of companies like Red Hat
and VA and the rest support the work of not-for-profit organizations or
efforts by "sponsoring" someone, who is then officially an employee of
that company, so this person gets a salary and benefits and all that. Then they
give this person the job description of focusing all of his or her time on the
not-for-profit activity. If someone were to want to support an
anti-software patent effort like that, would you be OK with it?


Richard M. Stallman: Of course.


LinuxWorld: You have talked with Tim O'Reilly about this issue.


Richard M. Stallman: I've talked with him to some extent. And I'm glad he managed to
persuade [Amazon CEO Jeff] Bezos to say that software patents need
changes. That may help.


LinuxWorld: Do you think that Amazon is moving in the right direction?


Richard M. Stallman: Well, Amazon is not moving at all. Basically, Bezos says that he
advocates a change in the law, and that would be a good change. But of
course a change in the law has to be made by Congress, not by Bezos. And
as far as Amazon's own policy is concerned, it hasn't changed.


So I agree with Bezos that [a change in the] law would be a good thing.
But meanwhile, I criticize Amazon's conduct, which is the same as it was
before.


LinuxWorld: Would you be satisfied with the extent of the changes he is
proposing?


Richard M. Stallman: Yes, I think that we could live with three-year patents or even five-year
patents. Five-year patents might be harmful, but at least not disastrous.


LinuxWorld: I have to admit that I'm really surprised to hear you say that! Don't
even short-term patents impede the freedom of programmers?


Richard M. Stallman: Yes -- but the magnitude of the impediment would be much less.
Short-term software patents would not eliminate the problem, but they
would reduce it to a point where we could cope.


LinuxWorld: Why did you call for a boycott of Amazon, and not of Priceline --
which is also guilty of this kind of behavior?


Richard M. Stallman: Actually, I don't know that much about Priceline. What happened was,
Don Marti [of the Silicon Valley Linux User's Group] proposed to me that
Amazon could be a good example to boycott for a couple of reasons. One,
they sell to the public; and two, they are actually suing somebody, which
makes them an egregious offender. There are lots of companies that have
software patents, and there are lots of companies that have software
patents on ideas that are more or less obvious and trivial, but only a few
have gone to the point of actually suing.


LinuxWorld: I think Priceline did sue. They sued Microsoft, I believe. [At the
time that this article went live, the case has still not been resolved.
--Eds]


Richard M. Stallman: Maybe. I don't know, actually... I don't know what Priceline does. Do
they do business with the public?


LinuxWorld: Yes, it's one of those online auction houses, I think.


Richard M. Stallman: It might be that that would be a good idea, too. I just haven't
thought about it.


LinuxWorld: : Would it be possible to say that you advocate a boycott of any company
which uses its patents aggressively?


Richard M. Stallman: A boycott would be legitimate against any company that uses a
software patent for aggression, but whether and whom to boycott is a
strategic choice. Often it is better to boycott just some of the
offenders rather than all at once.


LinuxWorld: In one of your texts about patents at the Free Software Foundation
site, you recommend that developers not get patents for their
innovations. Would you still recommend that?


Richard M. Stallman: Yes. I think that [not getting software patents] is one good thing to
do. The other good thing to do would be, if you are going to get a patent,
to make a binding commitment to use it only for defense -- that is, either
to push for cross-licensing or to put it in a patent pool.


LinuxWorld: And if someone were interested in helping to further the cause of
creating this kind of patent pool, what would you suggest that they do?


Richard M. Stallman: I'd suggest that the crucial first step is to find some companies
[which] have some patents that would be interested in joining the pool.
Because part of the reason it isn't terribly useful to design the details
of a pool in advance is that, since the crucial hurdle is to get people
with some patents to participate, it's really important to make sure that
the pool is designed in a way that appeals to them.


LinuxWorld: So one of the problems is to get somebody who actually has the patents
to be willing to be the first to give them away.


Richard M. Stallman: Yes. When you get some people together who have some patents and want
to join a pool of this sort, then talking with them, you can work out the
details, such that they will be willing to join it.


LinuxWorld: So, if it were solely in the interest of launching a patent pool, it
might actually be a good idea for people who oppose patents to go out and
get a lot of patents, just to be able to put them in the pool?


Richard M. Stallman: Yes, if you can be sure the patents won't be used for aggression
instead.


LinuxWorld: What about companies like Red Hat and Caldera and Corel -- they seem
to be in a unique position to take some responsibility in this sort of
project. Many of them have funding, so they can pay for the patents -- and
they believe in freedom, so they wouldn't mind starting the pool. What do
you think?


Richard M. Stallman: I hope they will. I would trust Red Hat to use patents for defense.


Corel and Caldera, by contrast, might use them for aggression instead. Of
course, I would be glad to see them show that they are better than I
realize.


LinuxWorld: The details of the pool seem kind of complicated, too. You
can't just make them public-domain patents, because then you have no
advantage over the others.


Richard M. Stallman: Right. You have to keep them active. And part of the problem is that
it costs money to keep a patent active. And it can cost a lot of money. So
people are torn between the idea of somehow licensing the patents to
non-members for money, and thus getting the money to apply for patents and
keep the patents going, versus the idea that you need to put pressure for
people to join the pool.


And the way you put pressure on other companies to join the pool is by not
letting them license these patents other than by joining. So it's
difficult.


And a further issue is that a patent pool or any other use of patents for
defense is not a totally effective approach.


It works when you are attacked by a patent wielded by a company that
really makes products, because their products may infringe some of the
patents in the pool. But some patents belong to money-squeezers, to
companies whose sole business is squeezing money out of people using
patents. And they don't have any products, so they are not vulnerable to
the pressure to cross-license.


LinuxWorld: Would you say that these IP trusts are an illustration of exactly how
bad an idea "intellectual property" is?


Richard M. Stallman: I'd say that the concept of "intellectual property" is too large a
generalization. It's not a useful concept to think about. It's very
important to distinguish between copyrights and patents and trademarks,
because they are totally different, both in how the laws work, and in
their effects.


So to generalize with a concept like "intellectual property" that includes
all three is simply asking for error. So you should throw away the term of
"intellectual property" and not find another term to replace it. Don't try
to talk about more than one of those different systems at a time, because
they are too different. If you try to talk about several of them at once,
you will be making mostly mistakes.


LinuxWorld: So if companies weren't using patents to protect their software, do
you think that...


Richard M. Stallman: Well, I object to the term "protect." I don't think patents protect
anything; patents are a method of attacking. They are a weapon of attack.
They endanger others. They don't protect anyone.


LinuxWorld: Sorry. I was thinking more in terms of the Constitution, you know, "To
promote the..." well, I can't remember the exact wording.


Richard M. Stallman: Right. The Constitution authorizes the existence of a patent system
as a method of promoting progress. But there is a lot of experience with
the software field from before there were patents, and there is no reason
to believe that you need patents in software to have progress. In fact,
there is no reason to believe patents promote progress in
software. Maybe in some other fields they do. I can see how, in the
absence of patents on drugs, it would be hard to get the money to pay for
the testing of them.


LinuxWorld: But we also had drugs before they were patented. It seems to me that
dedicated researchers are as willing to come up with new drugs as surgeons
are to come up with new surgical procedures?


Richard M. Stallman: I don't know enough to have an opinion on these questions about
patents and drug development. All I want to say about it is that it is
silly to insist arbitrarily that software and drugs must be treated alike.
There is no reason for "one size fits all."


But in the field of software, which is famous for the low barriers to
entry ... basically the big job is writing a large program, and what
makes it a big job is to combine a lot of existing ideas. Having ideas is
easy in software -- it's implementing these 3 new ideas with those 50
old ideas that's the big job. So as a by-product of doing these big jobs
over and over, they have new ideas.


The patent system encourages people to have ideas, and lock them up,
sequester them from use by others. And in this way, it obstructs the main
job of software development. Which means that it could give us fewer
ideas, because ideas just naturally come to people in the course of
software development. We don't need any artificial measures to encourage
people to have more ideas in software.


LinuxWorld: So the whole idea of protecting...


Richard M. Stallman: Well, I object to the term "protect," when what you really mean is
"monopolize." An idea is not going to be damaged by being used by more
people -- ideas don't require that kind of protection. So the term
"protect" embodies a number of biases and myths, and so I won't use it
for this.


No, I don't believe that the patent system in the field of software
promotes progress. We had plenty of progress before we had patents. There
is no reason to think that we have "more progress" now.


LinuxWorld: Is there anything else you'd like to say on the subject, but which I
forgot to ask about?


Richard M. Stallman: Europe is currently considering whether to allow software patents.
There is a proposal to amend the treaty that governs the European patent
office to remove a provision that currently prohibits patenting software.


Now is the time that the public can influence the decision. These are
political decisions, so the crucial thing is to look at how the public can
alter what is going to happen. There is a Website about this issue, which
is www.freepatents.org.


Now there is a chance in Europe to engage in political action to prevent
that change.


LinuxWorld: Thank you very much for the interview.


Resources



Richard Stallman on software patentsTim O'Reilly on software patentsJeff Bezos on software patentsReactions to Jeff Bezos' suggestions


LinuxWorld.com