Издательство Brookings Institution Press, 2008, -192 pp.
I am a quant. Almost all of my work involves either pure math or writing computer code to calculate numbers—even this book will rely on one mathematical theorem (which I promise will be painless). In my free time, I’ve done things like write a video game and a short guide to statistical analysis using the C programming language. But this is a policy book about writing good laws, so let me summarize what I want from the law: to be left alone to write code and do math. I am not trying to rationalize or alibi theft; I just want to be able to do calculations and write programs using my own ingenuity without worrying that anyone will have a claim on my results.
I am not exactly worried that a jackbooted lawyer might kick down my door at three in the moing. I distribute little of my code to the world, and statistics is not the most lucrative of fields. But I can’t write code in a vacuum and need tools written by others to do the hard parts. With increasing frequency, software projects are being shut down or crippled because of legal problems. As a result, I can do less.
When my brother bought a Linux PC so he could run simulations of new circuit designs, he asked me (his twenty-four-hour tech support hotline) how he could use the machine to listen to his music in MP3 format. Sorry, I explained, that would infringe patents. You’ll need to purchase a copy of Windows, since Microsoft has negotiated the appropriate licenses. What about using the PC to watch the DVDs in his living room? Nope, violates federal law. Can’t be done unless you want to be an outlaw and download C code from Brazil and recompile it on your PC. How about just watching QuickTime movies from the Inteet? Sorry again, but QuickTime depends on the Sorensen Codec, and Sorensen won’t let Linux programmers write anything that can read its data structures. Opening Word documents? You can do that, with OpenOffice.org, but there’s an application pending with the U.S. Patent and Trademark Office that may let Microsoft shut that down.
As for the electrical engineering simulations he’d originally bought the computer for? Those might be OK, but see page 63 for a list of patents that his simulations may inadvertently infringe if he calculates his Fourier transforms in certain manners. (His work frequently involves simulating integrated circuits, and he holds hardware patents 6,147,653 and 6,239,
755. However, he has never written a video game.)
Another time, he mentioned that he was thinking about writing a package to do simulations for nonlinear optics problems, since there are none on the market now. I’m safe using algorithms published in joual articles? he asked. The answer is once again no. On page 88, I discuss a data compression method from a peer-reviewed joual that led to patent headaches for those who were foolish enough to apply a published mathematical algorithm without consulting a lawyer first. In my own statistical work, I make use of factor analysis reasonably often, but some methods of factor analysis are now covered by patent 6,807,536.
An entirely new economic arrangement has appeared in mathematics and its offspring, computer science. Before, we were free to do whatever our abilities allowed, since mathematical and computational results were in the public domain—nobody could own an idea. This arrangement worked to bring us the mathematical and computationally advanced world we live in today. But in the past decade, a new set of rules has been imposed: an individual can own a mathematical result that he or she has discovered and can sue those who do not ask permission to use that result—even if the other person independently derived it.
The ownership of mathematical algorithms is truly a new concept and engenders one of the main questions underlying economics and law: what can a single human being claim ownership of? Although people sometimes describe property ownership as natural, it is clearly a social invention, designed to overcome economic and social problems. For example, the Earth was here 4.5 billion years before I was, and yet I am the sole owner of that piece of it that rests under my house. As a society, we have established this property right as a sensible solution to the sitcom-esque problems that would arise if anybody could show up at my home and use it as their own. Conversely, the benefits to private ownership of the sidewalk in front of my house are significantly fewer, so society has granted no one private ownership of it. There are other economic arguments for the ownership of abstractions like the design of a machine, which I discuss in this book.
The reader has no doubt been exposed to more than enough rhetoric about the fact that we live in an information age and our economic progress depends on the efficient movement and processing of information— and efficient information usage depends on better mathematical algorithms. But does inventing (and enforcing) the concept of ownership of a mathematical theorem make for a better economy? Is a mathematical algorithm more like a house or a sidewalk?
This is the first book to seriously ask whether it makes sense to allow for ownership of a computational algorithm. This question is not about the metaphysics of ownership, but about economic practicalities: because individuals can own the results of their research, they are more likely to innovate, but when you can’t use the math without permission, implementing and using the innovations become more costly. Since the new protections are not unambiguously a plus, we have to do the cost-benefit analysis to determine whether the new innovation they bring about is worth the trouble they cause.
Introduction
Optimal Breadth
From Equations to Software
Patenting Math
Profiting from Overbroad Patents
The Decentralized Software Market
Interoperability
Protecting Text
Policy Recommendations
I am a quant. Almost all of my work involves either pure math or writing computer code to calculate numbers—even this book will rely on one mathematical theorem (which I promise will be painless). In my free time, I’ve done things like write a video game and a short guide to statistical analysis using the C programming language. But this is a policy book about writing good laws, so let me summarize what I want from the law: to be left alone to write code and do math. I am not trying to rationalize or alibi theft; I just want to be able to do calculations and write programs using my own ingenuity without worrying that anyone will have a claim on my results.
I am not exactly worried that a jackbooted lawyer might kick down my door at three in the moing. I distribute little of my code to the world, and statistics is not the most lucrative of fields. But I can’t write code in a vacuum and need tools written by others to do the hard parts. With increasing frequency, software projects are being shut down or crippled because of legal problems. As a result, I can do less.
When my brother bought a Linux PC so he could run simulations of new circuit designs, he asked me (his twenty-four-hour tech support hotline) how he could use the machine to listen to his music in MP3 format. Sorry, I explained, that would infringe patents. You’ll need to purchase a copy of Windows, since Microsoft has negotiated the appropriate licenses. What about using the PC to watch the DVDs in his living room? Nope, violates federal law. Can’t be done unless you want to be an outlaw and download C code from Brazil and recompile it on your PC. How about just watching QuickTime movies from the Inteet? Sorry again, but QuickTime depends on the Sorensen Codec, and Sorensen won’t let Linux programmers write anything that can read its data structures. Opening Word documents? You can do that, with OpenOffice.org, but there’s an application pending with the U.S. Patent and Trademark Office that may let Microsoft shut that down.
As for the electrical engineering simulations he’d originally bought the computer for? Those might be OK, but see page 63 for a list of patents that his simulations may inadvertently infringe if he calculates his Fourier transforms in certain manners. (His work frequently involves simulating integrated circuits, and he holds hardware patents 6,147,653 and 6,239,
755. However, he has never written a video game.)
Another time, he mentioned that he was thinking about writing a package to do simulations for nonlinear optics problems, since there are none on the market now. I’m safe using algorithms published in joual articles? he asked. The answer is once again no. On page 88, I discuss a data compression method from a peer-reviewed joual that led to patent headaches for those who were foolish enough to apply a published mathematical algorithm without consulting a lawyer first. In my own statistical work, I make use of factor analysis reasonably often, but some methods of factor analysis are now covered by patent 6,807,536.
An entirely new economic arrangement has appeared in mathematics and its offspring, computer science. Before, we were free to do whatever our abilities allowed, since mathematical and computational results were in the public domain—nobody could own an idea. This arrangement worked to bring us the mathematical and computationally advanced world we live in today. But in the past decade, a new set of rules has been imposed: an individual can own a mathematical result that he or she has discovered and can sue those who do not ask permission to use that result—even if the other person independently derived it.
The ownership of mathematical algorithms is truly a new concept and engenders one of the main questions underlying economics and law: what can a single human being claim ownership of? Although people sometimes describe property ownership as natural, it is clearly a social invention, designed to overcome economic and social problems. For example, the Earth was here 4.5 billion years before I was, and yet I am the sole owner of that piece of it that rests under my house. As a society, we have established this property right as a sensible solution to the sitcom-esque problems that would arise if anybody could show up at my home and use it as their own. Conversely, the benefits to private ownership of the sidewalk in front of my house are significantly fewer, so society has granted no one private ownership of it. There are other economic arguments for the ownership of abstractions like the design of a machine, which I discuss in this book.
The reader has no doubt been exposed to more than enough rhetoric about the fact that we live in an information age and our economic progress depends on the efficient movement and processing of information— and efficient information usage depends on better mathematical algorithms. But does inventing (and enforcing) the concept of ownership of a mathematical theorem make for a better economy? Is a mathematical algorithm more like a house or a sidewalk?
This is the first book to seriously ask whether it makes sense to allow for ownership of a computational algorithm. This question is not about the metaphysics of ownership, but about economic practicalities: because individuals can own the results of their research, they are more likely to innovate, but when you can’t use the math without permission, implementing and using the innovations become more costly. Since the new protections are not unambiguously a plus, we have to do the cost-benefit analysis to determine whether the new innovation they bring about is worth the trouble they cause.
Introduction
Optimal Breadth
From Equations to Software
Patenting Math
Profiting from Overbroad Patents
The Decentralized Software Market
Interoperability
Protecting Text
Policy Recommendations