The patent balance
June 1st, 2011 at 11:19 pm by Dr. Drang
Marco Arment just posted a well-written argument against software patents, and while I am, as I said before, sympathetic to the notion that software patents should never have been allowed, I don’t think he’s taking the right approach and I can’t let his argument go by without comment. Duty calls.
Marco’s argument is perfectly pitched to make his case, especially to an audience of young programmers. It’s very persuasive writing unless you see how he’s stacking the deck.
The gist of his reasoning is in this passage:
Patents have a clear goal, as Dr. Drang quoted from the U.S. Constitution:
To promote the Progress of Science and useful Arts, by securing for limited Times to Authors and Inventors the exclusive Right to their respective Writings and Discoveries;
Excluding copyright (“Writings”), which I believe is a net positive, this implies that:
- Science would not advance sufficiently without inventors being able to legally prohibit others from copying or using their discoveries.
- The act of invention, not production or bringing to market, is where we want to place the most value as a society.
- The value we place on the act of invention is higher than the cost imposed on the rest of society by not being able to use the invention freely.
How could anyone disagree with this? Patents retard the advance of science, put inventors on a pedestal, and impose a high cost on the rest of society. Clearly, it’s time to break out the torches and pitchforks!
But before we do, let’s look at his three bullet points a bit more carefully.
First, patent law isn’t about science, it’s about invention, so patents aren’t going to restrain the advance of science. In addition, the prohibition that patents provide are limited in two ways:
- They’re limited in scope. Each patent has a claims section, and it is only the items covered in the claims section that an inventor is given dominion over.
- They’re limited in time. Currently that time is 20 years, which is, admittedly, an awfully long time in software-years, but it isn’t a permanent prohibition.
So by calling it science instead of invention and omitting the limitations, Marco’s made things sound much worse than they are—this is good polemic but not a good explanation of the patent system.
My biggest beef is with his second bullet point. Patent law does not put inventors at the head of society, it simply gives them some protection. Society already rewards those who bring products to market because they’re the ones who collect the money. Without patent protection, inventors would be cut out entirely because they’re not in the normal revenue stream. The purpose of patent law is to ensure that inventors get some reward for their efforts.
My disagreement with the third point is basically the same as my disagreement with the first. Society is able to use the invention freely, just not until the inventor has a chance to make some money off it.
In short, Article I, Section 8 doesn’t imply any of ideas Marco imputes to it. The purpose of intellectual property law is to strike a balance between the needs of creators and the needs of society as a whole. It will always be imperfect because these needs cannot be balanced exactly and because different people will view the balance differently.1 But just because it’s imperfect doesn’t mean it should be abolished.
None of the foregoing should be construed as enthusiastic, or even tepid, thumbs-up for the current software patent system.
Once upon a time, software wasn’t patentable. The thinking was that algorithms were basically math and you can’t patent math. Eventually there was a recognition that software does things, just like a traditional mechanical device. It seemed fundamentally wrong to allow patents on mechanical devices but not on software devices. Many programmers back then wanted software patents and worked hard to get them.
My sense is that most programmers would now argue against software patents, just as Marco has. We’re 30 years into the software patent system and seeing its downsides: the patent term is way too long for software; too many patents have been issued; and patent extortion is rewarded instead of punished. What’s gone wrong?
What’s gone wrong is that changes in the software landscape have turned the system I described above on its head. Software development requires essentially zero capital investment. With the rise of the personal computer and the Internet, software distribution also requires essentially zero capital investment. Programmers—the producers, the people who bring things to market—can now be the little guys. And through the accumulation of patent portfolios, “inventors” who don’t do any inventing at all can be the big guys.
Marco is right when he says patent law, when applied to software, is based on assumptions about what should be rather than what is. This, I think, is the crux of the software patent problem. Things are out of whack because we’ve taken a system designed for the Industrial Revolution and applied it to a post-industrial realm.2
The solution, I think, is a patent system that recognizes that software and hardware are not the same. The patent balance has always been based on economics—what rules and incentives are needed to get inventors to create the devices that improve society? Because the economics of software are different from that of hardware, the rules and incentives should be different, too. This is not unprecedented. All the various intellectual property laws have been tuned to their specific needs. Splitting software out from hardware would just be a recognition of its unique needs.
I said at the top that I’m sympathetic to the idea that software patents should never have been allowed. But whether they should have been allowed or not, they are a fact of life today and that’s not going to change. If you think the patent system is out of balance, you should work to put it back in balance, not chase after some fantasy of turning back the clock.
-
Marco, for all his dislike of patent law, gives a pass to copyright law, even though it does the same thing for a much longer term. He sees the copyright balance differently because he has no stake in it. ↩
-
My family saw Thor over the weekend, and realm has worked its way into my everyday vocabulary. Verily. ↩
June 2nd, 2011 at 9:13 am
“Society is able to use the invention freely, just not until the inventor has a chance to make some money off it.”
“freely” - I don’t think that word means what you think it means.
If producers are paying money to inventors (or, more likely, to lawyers and/or patent holding companies) then they are not using the inventions freely. They are using it at a (potentially large) cost.
June 2nd, 2011 at 9:17 am
And as soon as I hit post I realized that you used the words “not until”. So, um, nevermind.
Nonetheless, 20 years is an awfully long time and the costs can be extremely high.
It seems like the current system benefits lawyers much, much more than inventors or producers.
June 2nd, 2011 at 9:32 am
I’m sorry, but Apple, a group of developers, or someone needs to fight this all the way to the Supreme Court because software patents are a blatant double dipping of intellectual property. Software is copyrighted material thereby a “form” of software can’t be patented. That patents have been granted is unconstitutional because the U.S. Constitution makes a distinction between the two. Copyrights, as all software bears, apply to the specific presentation of a work. Patents apply to form. Giving a type of software a patent is like giving George Lucas the sole right to create a film script or Steven King the sole right to publish novels.
June 2nd, 2011 at 1:00 pm
I think you guys didn’t understand what he meant with this post. He’s not saying that the current patent system is suitable for software, but that patents are necessary.
Society will use and benefit from that invention, and producers will happily pay for it as long as they make some benefit from using the invention in question.
And if you knew a bit about economy, you’d instantly agree that patents foster innovation. There are even maths on that, that prove it. And math is the ultimate argument, guys. That 20 years is a long time for the particular case of software is a fair point, but abolishing the patent system altogether would be an abomination in the long run, both for consumers and inventors.
Mark my words^^
June 2nd, 2011 at 1:16 pm
Jay,
Thanks for mentioning the dual copyright/patent nature of software. I wanted to say something about that in the post but forgot. To me, that’s another example of how software is different and needs different rules.
June 9th, 2011 at 5:26 am
Why? Why do we have to live with software patents? You basically conceded here that patents on software have turned out to be a terrible idea and do much more harm to society than good, and we’d all be much better off without them. But then you turn around and say that there is no turning back the clock and we just have to learn to live with them. To which I respond - why?
There are plenty of countries where you can’t patent software. New Zealand for example has recently explicitly removed software from patentable subject matter. So it clearly can be done. Or are you just saying it can’t be done in the US? Why not in the US? It can be done in other countries? Are you saying that the US government is too dysfunctional corrupt and broken to be able to implement what everyone agrees is a good idea? I can’t think of what else you could be saying really. It seems like an awfully negative point of view to have about your own government.
June 9th, 2011 at 6:20 am
“they are a fact of life today and that’s not going to change.” and yet software patents in the US changed just last year with the Supreme court’s Bilsky decision. Unfortunately that decision was pretty unclear. It probably invalidated a lot of US software patents. It may have invalidated them all. This can and do change.
June 9th, 2011 at 10:56 am
Software patents should not exist.
All software is nothing but mathematical expressions that describe something. In and of itself, software is nothing but an abstract idea, expressed in a rigid mathematical notation suitable for automatic interpretation by a machine.
Only devices that translate the inputs and outputs of that machine into physical actions should be patentable.
June 9th, 2011 at 11:05 am
“Once upon a time, software wasn’t patentable. The thinking was that algorithms were basically math and you can’t patent math. “
That’s correct.
“Eventually there was a recognition that software does things, just like a traditional mechanical device. It seemed fundamentally wrong to allow patents on mechanical devices but not on software devices. Many programmers back then wanted software patents and worked hard to get them.”
Whatever, but software is still basically math and you can’t patent math. Except for in the USA and Japan, apparently.
Also if you remember the shit-storm Unisys invoked over themselves when they decided to patent and charge money for the LZW algorithm in GIF pictures, I’m not so sure about your claim “Many programmers back then wanted software patents and worked hard to get them.” I think it was more the programmers’ corporation’s lawyers :-)
In the words of Bill Gates in 1991 (before software patents became commonplace): “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.… The solution to this is patent exchanges with large companies and patenting as much as we can. “
That doesn’t sound very innovation-enhancing to me. Unless you quantify innovation as numbers of patents issued, of course.
June 9th, 2011 at 2:25 pm
Ian,
To paraphrase the late great Emily Litella, “You ask a lot of questions for someone from New Zealand.”
The US and NZ operate under different sets of rules, and the Supreme Court has already ruled that the US Constitution allows software patents. Now, it’s true that SCOTUS decisions get overturned from time to time, but it’s nuts to think that the current Court will overturn a ruling that’s supported by large corporations.
And by the way, I do think the plain language of the Constitution supports software patents. As I said in the post, that doesn’t mean I think the current patent rules are a good fit for software, but I do think a better and workable set of patent rules is possible.
One need only look at the way copyright has been extended time and again—basically, any time “Steamboat Willie” is about to slip into the public domain—to see how difficult it is to get IP reform. It’s hard to be optimistic in the face of these facts, but I still hold out hope that abuses of the patent system, and the outrage they cause, will lead to improvements.
June 9th, 2011 at 2:34 pm
Jesse and Frits,
I’m surprised to hear the “software is just math” argument still being used in 2011. Software does things, math just is.
It’s true that software doesn’t do physical things (and that is, by the way, the reason my teeth grind when I hear programmers called “engineers”), but I don’t see how that excludes it from what the Constitution calls “the Useful Arts.”
June 9th, 2011 at 5:29 pm
“I’m surprised to hear the “software is just math” argument still being used in 2011.”
Huh? Why are you surprised that it’s “still used”? Arguments aren’t like transistors or ball bearings, they don’t wear out. It’s precisely because they are abstract that they don’t wear out; they’re very much like software in that way.
“Software does things, math just is.”
Actually, software doesn’t do things, it is a set of instructions for doing things. The doing part is what the computer is there for.
June 9th, 2011 at 7:43 pm
mike, I don’t think “wear out” is the best description, but arguments do get abandoned because they stop working. The world changes, and arguments that were once persuasive cease to be so. Few people would argue today that blacks are inherently inferior or that women are temperamentally unsuited for positions of authority. These are, however, arguments that were both common and persuasive 50+ years ago.
Don’t like examples from sociology? Let’s look at the natural sciences, then. Arguments based on the ether or phlogiston aren’t very common anymore, but these were once abstractions that had wide currency.
So, neither arguments nor abstractions last forever. If you want to argue that software is simply math—and therefore not patentable—be my guest, but that argument hasn’t worked in 30 years.
Your final paragraph is simply tendentious. You could just as easily say the pull-tab on the top of a Coke can doesn’t do anything itself; the doing part is what the person’s finger is for. But of course the person’s finger wouldn’t be able to open the can without the clever (and patented) pull-tab configuration there to transform the finger’s force and motion into something useful. By the same token, software installed in a computer configures the computer so that the electricity flowing through it is transformed into something useful.
June 11th, 2011 at 10:10 am
“Once upon a time, software wasn’t patentable. The thinking was that algorithms were basically math and you can’t patent math. Eventually there was a recognition that software does things, just like a traditional mechanical device.”
This is wrong. Software does not “do” anything; software is data accessed by a device, and it is the device which “does things”.
June 11th, 2011 at 1:48 pm
Saul,
Assuming anyone other than me reads your comment, it will be after reading Comments 12 and 13, and they’ll wonder why it is you bothered to parrot an assertion already made by someone else and didn’t address my response to that assertion. They’ll probably conclude that you’re lazy and not too bright, which is probably not the impression you wanted to give.
June 12th, 2011 at 12:19 am
I am much more comfortable with the accusation of committing the offense of reiterating a truism than I’d be with the thought of being the pioneer in expressing a falsehood. I lay no exclusive claim to the idea expressed, nor would I expect those who expressed it previously should demand such a monopoly on “their” idea (wouldn’t that be ironic!?).
In response to your pop-top analogy, I guess you are saying that after somebody has obtained a patent upon the pop-top/lever design, another person should be able to later patent the (software) process of inserting a finger under the tab and lifting? If so, I should strongly disagree: the second patent should be disallowed because it is merely specifying how to use the device as it was designed to perform — just as any software is merely the specification of how a computer should perform the actions it was originally designed to perform.
June 12th, 2011 at 11:01 am
Saul, Saul, Saul,
June 12th, 2011 at 4:45 pm
I’m not sure there is much of a difference between a patent on software and “composition of matter” patents. One can patent the formula for Coke or a motor oil which is simply a set of instructions that by themselves do nothing. Some other entity has to manifest them by measuring the components and mixing them. How is this different from a software set of instructions that when implemented do something useful?
A point about patents that the article failed to make is that in return for a limited time monopoly, the inventor must teach the invention in the patent specification. One could argue that making inventions public like this furthers progress more than keeping inventions as trade-secrets, never to be revealed.
June 12th, 2011 at 11:55 pm
Yosh,
Yes, full disclosure is an important part of the balance, and I should have mentioned it in the post. Full disclosure, along with the time limit, is the reason Coke’s formula isn’t patented—Coke wouldn’t be able to control it if it were.