" The Congress shall have Power …
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;"
United States Constitution, Article I, Section 8
The clause in the Constitution that covers intellectual property mentions two categories of protectable creations: writings and discoveries. It says nothing about differences in how the two sorts of creations are to be protected. Yet United States law, following earlier English law, provides two entirely different systems of protection: Copyright law for writings, patent law for inventions. In this chapter we will try to make sense of the differences in the two bodies of law and the allocation of the various sorts of intellectual creation to the different categories, using ideas worked out in chapter 10. The first step is a brief sketch of copyright law, patent law, and how they differ.
"… where both ... central characters have miraculous strength and speed; conceal their strength, along with their skin-tight acrobatic costumes, beneath their ordinary clothing; are termed champions of the oppressed; crush guns; stop bullets; and leap over or from buildings."
Detective Comics, Inc. v. Bruns Publications
Copyright was originally available only for writings but has since expanded to cover computer programs, pictures, music, and much else. What it controls is the right to make copies (and derivative works, such as translations). This raises the question of exactly what feature of the work it is that you are not permitted to copy. The standard answer is that copyright law protects expression but not idea. If I copy a chapter from your novel, I have violated your copyright. If I instead copy the plot, the idea of your novel, while providing my own characters, incidents, and background, your copyright has not been violated. Patent law, in contrast, protects idea not expression. What you patent is not a particular machine but a set of ideas that are embodied in that machine and could be embodied in other machines.
The distinction between idea and expression blurs on close study. Is a superhero who can see through walls, is impervious to bullets, and jumps over tall buildings at a single bound an idea, expressed in a particular Superman comic, or an expression of the idea of a superhuman being?
That particular question has been settled by the courts; Superman won an infringement suit against an imitator who had borrowed too many of his characteristics. The issue remains unsettled in the newer and more important context of software copyrights. Is the Lotus 123 menu tree, the particular pattern of commands used by what was for some years the dominant spreadsheet, an idea, expressed in the form of a particular program, or an expression of the idea of using key commands to make a spreadsheet do things?
That became an important legal issue some years back when several software companies produced clones of Lotus 123. Their programs, written from scratch without copying any Lotus code, were designed to work like Lotus, using the same commands in the same pattern so as to make it easy for customers to switch over without retraining. The result of extended litigation was ambiguous, with some courts taking one side, some the other; the one decision that reached the Supreme Court was permitted to stand on a 4/4 tie vote. While it is clear that copying someone else's computer code or the precise appearance of the screens it produces is a violation of copyright, it is unclear at what point beyond such literal copying protectable expression turns into unprotectable idea.
One restriction of copyright is that it applies only to expression, not idea. A second is that it only applies to copies, not independent creations, even if they happen to be identical to the copyrighted material. If I independently write a chapter that reads exactly like your chapter and can prove it, my chapter is not covered by your copyright.
The probability that two authors will independently write the same chapter is not very high. But consider the analogous case in computer law. In order to clone the IBM PC, to make computers that would run the same software in the same way, it was necessary to produce functional equivalents for the ROM chips built into the original. Those chips contained firmware, bits of computer code that the computer used for various basic functions. Computer code is covered by copyright. The companies designing the chipsets that would make cloning possible had to find some way of reproducing the function of the chips without copying them.
The solution was to set up two teams of engineers. Team A examined the IBM chips, read the computer instructions built into them and figured out, in agonizing detail, exactly what they did. The result was a set of functional specifications: If you put this number in this memory location, this is what happens. When the specifications were complete, describing what the engineers believed to be all of the functions of the chip, they were handed to team B.
Members of team B, the "clean room" team, were never permitted to look at either the original chips or the code they contained. Their function was to reverse engineer the chip. They had to write code that would do exactly what the IBM code did, based on the specifications produced by team A, without seeing or copying the original.
If you set out to duplicate the detailed functionality of a lot of short bits of computer code, including the exact memory locations that numbers have to be put into or read out of, the code you produce is likely to look very much like the code in the original. It might even be identical. But for the purpose of the copyright law, that doesn't matter. Copyright law only protects against copying.
For me to write the same chapter you wrote is quite unlikely; if we wanted to wait until it happened by pure chance, we would need a good supply of candles, since the sun is unlikely to last that long. And if it did happen by accident, perhaps with something much shorter than a chapter, it would be very hard to prove. For the clean room team to end up recreating from the specifications large chunks of the original code is very much more likely, which is why their employers took great care to make sure they could prove in court that what they came up with was independently created, not copied. They were copying the function of the code, but function is idea not expression, hence not protected by copyright.
Independent invention is one loophole in copyright. Another is the doctrine of Scenes a faire, which holds that if there is only one way of expressing an idea, or only a very small number of ways, copying that expression is not a violation of copyright. Without that, a copyright on the expression would implicitly include ownership of the idea.
Another loophole is the doctrine of fair use, which sometimes permits copying of otherwise protected materials. The criteria are the nature of the use (non-profit and educational are good, commercial bad), the nature of the work, the amount copied (less is better), and the effect on the revenue of the copyright holder.
How do you get a copyright? The short answer is that you get a copyright by creating a protectable work. While the United States maintains a copyright registry, and registration of your work gives you some additional protection against infringers, it is not necessary to register in order to be covered by copyright. It is not even (now) necessary to attach a copyright notice.
How long does copyright last? In the United States, the term of protection has varied somewhat over the years, but always been long. Currently it is life of the author plus fifty years.
In summary, copyright is easily obtained, long lasting, clearly applies to literal copying of writings and a variety of other things, and may sometimes apply to less than literal copying.
Popular usage to the contrary, there is no such thing as a patented machine. Patents are not on objects but on ideas. To apply for a patent, you specify your claims, explaining precisely what ideas you claim to have invented, and describe the best way known to you of practicing your invention. If you have invented the electric telegraph, you describe both the ideas and the best way you know of to implement them—to build a telegraph.
Applying for a patent is the job of the inventer and his patent attorney. Denying it is the job of the patent examiner. It is up to him to determine whether the application meets the requirements for patentability, whether it is novel, nonobvious, and useful.
Novel means that your idea has not been thought up before, at least by anyone who published it, or publicly practiced the invention, or did other things from which it might have been learned. Nonobvious means that if someone else had wanted to do what you have done, it would not be sufficient for him to simply hire a competent engineer, someone "skilled in the relevant art", and tell him to do it. Something more, vaguely and traditionally seen as inspiration, a flash of genius, a light bulb going off over a cartoon head, is required.
Part of the fun of studying law is the stories, and one of my favorites has to do with the 19th century definition of useful. The case of Rickard v Du Bon arose when Rickard got a patent, Du Bon infringed, and Rickard sued. Du Bon, in his defense, argued that the patented process was useless, so the patent was invalid.
It seems a little odd for a company to go to the trouble of violating a patent on doing something not worth doing. It is even odder given the definition of "useful" in 19th century patent law. Courts held that if an invention was merely of no use to anyone, granting a patent on it did no harm. If the inventor thought it was useful and the patent examiner did not, the inventor was free to waste his own money getting a patent. To be unpatentable, an invention had to be not merely without use but positively pernicious. In the words of Judge Story in Lowell v. Lewis:
"All that the law requires is, that the invention should not be frivolous or injurious to the well-being, good policy, or sound morals of society. The word "useful," therefore, is incorporated in the act in contradistinction to mischievous or immoral. For instance, a new invention to poison people, or to promote debauchery, or to facilitate private assassination, is not a patentable invention."
So we have a case in which Du Bon is arguing, first, that Rickard's patented process is wicked hence should not be patentable, and second that since it is not patentable, he should be free to do it too.
The process was for artificially flecking tobacco leaves. Flecking had no effect on the quality of the tobacco, but flecked tobacco came from places that produced high quality tobacco, so cigar buyers interpreted a flecked appearance as a signal of high quantity. Rickard patented the process. Du Bon argued that the invention was pernicious, since its purpose was to deceive the customer, hence useless, hence unpatentable—so he should be allowed to do it too. He won the case.
Modern courts have strengthened the utility requirement, holding, for instance, that a process for producing a chemical that has no known uses is not useful, and thus not patentable.
In drawing up your patent application, you are defining your invention—stating just what ideas you claim to have invented. What makes the project difficult is that ideas do not come with clearly defined boundaries. The same invention can be described narrowly as the exact way in which a particular machine is designed or more broadly, in some cases very much more broadly, as the ideas whose implementation distinguishes that machine from all previous machines. It is up to you to draw the boundary around your invention, to stake out your claim in idea-space. It is rather as if a homesteader, instead of being limited to 160 acres, was allowed to decide for himself just how much land he had settled.
In idea space as in geographical space, more is better if you can get it. The problem is that if your claim is broadly defined, it is quite likely that something someone else has already done fits its description or that some way of doing something that fits the description is obvious. If, after inventing the world's first nuclear reactor, you describe it in your patent application as a way of getting power out of matter and claim all ways of doing so, the patent examiner can point out that the steam engine did that quite some time ago.
The narrower the claim, the less likely that problem is to occur. Continuing our geographical analogy, the smaller the territory you claim to have discovered, the less likely it is that someone else can show he got to some part of it first. The problem with making only a narrow claim is that it may not be worth much; next year someone else may invent around your patent, creating a reactor not covered by a narrow description of your invention.
When Samuel Morse drew up the patent application for his telegraph, he had a solution to that problem. His application contained six different claims. The sixth was for all ways of using the electromagnetic force to transmit letters or symbols to a distance. Not only did that include all telegraphs, it covered fax machines, television, and the Internet as well, despite the fact that none of them had been invented yet. The Supreme Court, in O'Reilly v. Morse, rejected that particular claim on grounds of overbreadth, while granting the five narrower claims that Morse (prudently) also included. They should have rejected it on the grounds that it was not novel. Signal flags, semaphores and beacon fires had all been using the electromagnetic spectrum, in the form of visible light, to transmit messages since long before Morse invented the telegraph.
As this description of the patent process suggests, one of its functions is to convert a commons in idea space, a commons consisting of all of the ideas that might be invented, into private property, with each inventor drawing up the lines of his particular claim. One scholar, Edmund Kitch, has argued that this is its principle function. His prospect theory of patents analogizes them to mineral claims. One reason for a mineral claim is to give prospectors an incentive to go out into the desert looking for gold, just as the possibility of getting a patent gives inventors an incentive to make inventions. But a second reason is that in order for the successful prospector to exploit his find he needs mineral rights to the land around it, enough land so that someone else cannot come in, dig a hole next to his, and free ride off his discovery. Similarly, in Kitch's view, ownership of a substantial chunk of idea space permits the patent holder to control and coordinate the further developments of his invention.
As we have seen, patent differs from copyright in a number of dimensions. Copyright applies to expression, patent to ideas. Copyright is obtained automatically, patent only after a laborious application process that many applications cannot survive. Finally, a copyright is good for the author's life plus fifty years; a patent, in United States law, has generally lasted for between fourteen and twenty-one years; current patents (with some exceptions) are for seventeen years.
In the previous chapter I sketched out a framework for determining what sorts of property rights should exist for what sorts of things. We have spent the first half of this chapter examining, in the context of intellectual property, what sorts of property rights do exist for what sort of things. It is time to put the two sketches together.
In considering the cost of a system of property rights, I started with the problem of drawing boundaries. That problem is minor in the case of copyright protection against literal copying. My chapter is or is not a copy of yours. Your copyright over your book is a sharply defined property right.
The same is not true of patent protection. Patent attorneys and patent examiners spend enormous amounts of time and effort haggling over just what should or should not be included in a particular claim. After they are finished and a patent has issued, attorneys for the patent holder and the putative infringer spend more time and energy, both their own and that of judges and juries, determining whether a particular machine or industrial process embodies ideas that are on the protected side of the line. Even after marking out his claim in voluminous detail, the inventor has a less clear idea of what he owns than the author, who has not bothered to mark out his claim at all.
A second and closely related problem is enforcement. Major violations of copyright take the form of publications, publicly observable. Because the right is well defined, both the litigation cost to copyright owners of proving that a violation has occurred and the cost to potential trespassers of avoiding an unintentional violation are usually low. Violations of patent take the form of machines or industrial processes, the details of which are sometimes secret. And even if everything is known, the fuzzy nature of the boundary around the property makes litigation costly and uncertain and accidental trespass, or mistaken charges of trespass that did not happen, almost inevitable.
A third problem is the transaction cost of using the market to move property to its highest valued use. Here again, propertizing expression makes more sense than propertizing ideas. The only way I am likely to copy your copyrighted writing is deliberately, in which case I know whom to apply to for a license. But an inventor risks reinventing bits and pieces of a dozen patented inventions. Avoiding a patent violation may require an extensive patent search, followed by license negotiations for permissions that may or may not be really necessary.
Finally, and perhaps most interestingly, is the issue of incentives to create rights. Providing an incentive to write books or make inventions is one obvious purpose of intellectual property law. What we want however, as I have pointed out before, is not merely an incentive but the right incentive. The gain to the producer of intellectual property should be at least a rough measure of the value to other people of what he produces in order that he will bear the costs of producing it if and only if it is, on net, worth producing.
When I write a book and enforce my copyright on it, I make it difficult for another author to write the same book; although independent creation is a defense against the charge of infringement, it is, in practice, difficult to prove. I have taken a tiny bit of property out of the commons: the opportunity to write that book. But the number of books that can be written is so enormously larger than the number that will be written as to make what I have taken worthless to anyone but me—nobody else was going to write that book.
It follows that if I write a book, copyright it, and collect royalties, the royalties I collect provide a conservative measure of the value I have produced. It is a conservative measure because, despite the best efforts of my publishers, some readers end up getting the book for less than they would be willing to pay for it. I can ignore the effect of my copyright on the opportunities available to other writers. They lose nothing through not being permitted to write my book, since they wouldn't have written it anyway.
The same is not true when I make an invention. The opportunity to invent something is often quite valuable, as demonstrated by patent races in which two or more teams are competing to make and patent the same invention. The first team to make the invention gets the patent and collects seventeen years of royalties. But if the second team would have made the invention six months later, the social value of the first team's work is six months of the invention, not seventeen years. The commons that the inventor depletes is very much more depletable than the commons depleted by the writer. It follows that the gain to the inventor may in part consist of a transfer from others who might have invented the same idea a little later. So patent royalties may over reward the inventor, providing too much incentive to make and patent new invention. If so, invention becomes in part rent seeking, an expenditure of resources by one actor in order to obtain benefits at the expense of another.
Putting all these arguments together, we get a simple explanation for the observed laws. Copyright protection against literal copying creates a form of property that is easy to define, cheap to enforce, relatively easy to transact over, and subject to no rent seeking problem. Hence we give copyright easily and for a long term. Patent protection creates a form of property that is hard to define, hard to enforce, costly to transact over, and contains a potential inefficiency due to patent races leading to duplication and inefficiently early inventions. Hence we give patents grudgingly and for a short term.
The same analysis explains at least some of the details of intellectual property law. The Mise en Scene exception to the general rule that expression is copyrightable covers the special case where depletion of the commons is a problem—an expression that is the only way (or one of the few ways) of expressing a particular idea. The patent requirements of novelty and nonobviousness are designed to control the problem of rent seeking by limiting patents to those inventions major enough so that there is a reasonable chance they would not have been made independently anytime early in the patent period. They do so very imperfectly, in part because patent is, with a few narrow exceptions, a one size fits all system; ideas come in a continuous range of size and durability.
Another detail that makes sense is the fair use exception. The smaller the amount you are copying, the greater the transaction cost of arranging permission relative to the value of the permission, hence the stronger the argument for treating the property as a commons. The smaller the effect of your copying on the revenue of the copyright holder, the smaller the reduction in his incentive to produce the work due to his inability to prohibit your use. Even the preference for non-profit and educational use may make sense, if we believe that those activities produce, on average, positive externalities, and it is thus efficient to subsidize them—in this case at the expense of the copyright holder, in other contexts through tax exemptions and direct government spending.
Consider, finally, the decision by modern courts to put content into the requirement that an invention must be useful in order to be patentable. Given the nature and speed of modern research, this requirement may serve a useful function. Without it, it would be in the interest of companies to search for ways of synthesizing chemicals for which no use was known in the hope that someone would eventually come up with a use and have to go to them for the right to use their synthesis. Since information on how to synthesize a chemical is only useful after some use for the chemical has been found, it is arguably more efficient to put off research on synthesis until you know whether or not the chemical is worth synthesizing.
While this is a plausible argument, it may not be a correct one. One possible response is that whether it is worth looking for a use for a chemical depends in part on how easy it is to make it, so there is some advantage to finding the synthesis first. Another response, based on Kitch's prospect theory of patents, is that by patenting the synthesis the inventor gets de facto ownership of the chemical, at least until his patent expires or someone discovers a different synthesis, giving him both the incentive and the ability to promote and coordinate the search for uses.
Another application of the economic analysis of property is to the question of whether computer programs are writings, and so protectable by copyright. In the early cases, judges expressed considerable uncertainty and reached differing conclusions. Some programs, after all, such as machine language programs burned into ROM chips of a computer, are never intended for human eyes. Such a program is, as John Hersey, a prominent author and the head of the Patent and Trademark Office, argued, more analogous to an elaborate cam, a part of a machine used to control other parts, than to a novel. As some early courts noted, the closest previous case involved the paper tape used to control a player piano, which had been found not to be covered by copyright—because it was not intended to be read and so was not a writing.
Arguing by analogy, Hersey and the judges who rejected copyright for (at least) machine language programs were right. Arguing by function, they were wrong.
The first step of the functional argument is to ask, not whether programs are writings, but why writings are covered by the particular sort of legal rules we call copyright. The previous section of this chapter offered a brief answer. The next step is to ask whether computer programs share the characteristics that make copyright appropriate for writings. The answer, for protection against literal copying, is yes. One program either does or does not contain a substantial amount of code copied from another, and it is usually easy to tell which is the case. Independent creation is unlikely, save in the special case of reverse engineering, where clean room techniques make it possible to prove independent creation. And a bar against literal copying of a particular program imposes an insignificant cost on later programmers, since they could not have copied the program if it had not been first written by someone else and are unlikely to have independently written exactly the same code by accident.
In discussing copyright, I have repeatedly limited myself to the case of protection against literal copying. When we go beyond that, when we try to use copyright to protect the Lotus menu tree or the look and feel of Apples' Macintosh interface, the arguments for copyright protection become much weaker. Such extensions make the protected property more like the ideas traditionally protected by patent law. That suggests that we might be better off limiting copyright law to literal copying and using something more like patent law to protect less clearly defined rights.
Copyrightability of programs is one example of a defensible extension of the legal rights designed to protect writing to other sorts of intellectual property. Another example is a state plug mold statute. Such laws prohibited a firm making (say) a boat hull from getting its design by buying a competitor's hull, forming a mold around it, and then using that mold to produce its own hulls. Here again, the argument in favor is not that boat hulls are writings but that this particular form of copying has the same features that make copyright law appropriate for controlling the copying of writings.
I referred to state plug mold statutes in the past tense because they have been held unconstitutional by the Supreme Court, presumably on the theory that intellectual property protection is the business of Congress, not the states. Somewhat similar protection is provided at the federal level for computer chips under the Mask Works Act and for boat hull designs under the Vessel Hull Protection Act (title V of the recent Digital Millenium Copyright Act).
Property rights serve two related functions: They provide both a way of deciding who gets to use what when and an incentive for creating things.
In the case of intellectual property, the first function is not merely unnecessary but perverse. We cannot both drive the same car to different places, which is an argument for property rights in cars. But we can both use the same idea to build different machines, or simultaneously read different copies of the same book, which is an argument against property rights in ideas or writings.
Having one more person choose to read a book does not increase the cost of writing it; that must be paid in full in order for there to be any readers at all. Hence when we include in the price of a book a royalty payment to the author, we are, from the standpoint of efficiency, overcharging. If the price of an eleven dollar book represents a ten dollar production cost plus a dollar in royalties to the author, an additional copy produces a net benefit as long as the purchaser values it at more than ten dollars. But if he values it at more than ten and less than eleven, he will not buy it, which is an inefficient outcome. The same is true for the potential user of an idea who values it at more than zero but less than the license fee set by the patentholder. On the dimension of how many people use an idea or a writing private property gives an inefficiently low result, commons an efficiently high one.
There remains the second function of property. It is hard to read a book if nobody has written it, and authors may choose not to write books if they cannot collect royalties on them. Similarly for inventions. So the protection of intellectual property does provide some benefit. But the case for treating ideas as property is, economically speaking, weaker than the case for propertizing many other things, which may help explain why intellectual property is a relatively recent institution.
"'Trade secret' means information, including a formula, pattern, compilation, program, device, method, technique or process, that: (1) derives independent economic value, actual or potential from not being generally known to, and not being readily ascertainable by proper means by, other persons who can obtain economic value from its disclosure or use, and (ii) is the subject of efforts that are reasonable under the circumstances to maintain its secrecy."
(Uniform Trade Secrets Act)
A trade secret is information whose value to its possessor depends on its not becoming generally known. Trade secret law defines the rights that the rightful possessor of a trade secret has against those who have obtained the secret wrongfully, through breach of contract, say, or an employee's violation of his duty of loyalty, or a trespass. It permits him to collect damages based on his loss or the misappropriator's gain. Sometimes, but not always, it permits him to enjoin use of the secret by someone who received it from the guilty party. An injunction will ordinarily not be available against a third party who could not be expected to have known that the trade secret was stolen and has taken actions, such as building a factory using the secret, that would leave him worse off if enjoined than if he had never received the secret.
Unlike patent or copyright, trade secret law does not provide protection against someone who has obtained the secret without the consent of the owner but without violating his legal rights, for example by reverse engineering a product that embodies the secret. It thus gives protection well short of full scale property rights.
This brief description suggests two puzzles. The first is why trade secrets are not, like patentable inventions, treated as property. The second is why, if trade secrets are not patentable, they are protected at all, why state trade secret law is not, like state plug mold statutes, preempted by federal patent law.
A legal regime in which trade secrets were entitled to property protection, in which, for example, a competitor who deduced a trade secret, or discovered it by reverse engineering, or invented it independently, could be enjoined from using it, would be a regime of self-claimed patents without examination. Such a regime would be inefficient for the same reasons as a regime in which the patent office granted patents to the first inventor of every idea. There would be an incentive to waste resources discovering and claiming ideas long before they were of any use to anyone. The incentive to appropriate is in this respect too high, since some of the gains represent transfers to the first inventor from people who could have invented the idea themselves and used it for free, but must now pay him a license fee. Thus the ability to freely appropriate property rights in ideas is a potential source of inefficient rent seeking.
Rent seeking appears a second time in the protection of property—as expenditure on prosecuting and defending infringement suits. Such expenditures would be eliminated if the ideas in question were entirely unprotectable. Whether this is an argument against intellectual property protection depends on whether the costs of legal protection, as under a patent system, are high or lower than the cost of protecting legally unprotected ideas in other ways, most notably secrecy.
One argument for the sort of limited protection that trade secret law provides is that, by making it easier to protect trade secrets, it makes them more productive. Without trade secret law, the possessor of a trade secret might be limited to using it in his own factory under his own eye, where he can make sure nobody is stealing it; with legal protection, he has the option of licensing it to other people, suing them for misappropriation if they steal it, and enjoining third parties to whom they sell it .
Whether rent seeking provides, on net, an argument for or against legal protection, it is at least an argument in favor of defining protectable rights by criteria that are easy for courts to measure. That may help to explain why trade secrets are not protected against reverse engineering; in most cases there is only a fuzzy line distinguishing reverse engineering, designed to discover a competitor's secrets, from research designed to produce better products. More generally, the requirement that the appropriation of a trade secret must involve some independently wrongful act in order to count as misappropriation provides a convenient bright line rule, one that takes advantage of the existing law that defines the relevant sorts of wrongful acts. Here as elsewhere, bright line rules have the advantage of reducing litigation costs.
This brings us to our second question: why trade secret law is not entirely preempted by patent law. To answer it, consider an invention that is unpatentable because the patent office considers it too obvious. Such an invention may not deserve 17 years of protection, but that does not mean that it deserves no protection at all. Trade secret protection provides some reward for the inventor, and thus some incentive to make the invention, without imposing any cost on others who might independently invent it. It thus helps fill one gap in the patent law, the lack of protection for inventions that are only moderately non-obvious.
Trade secret law also helps fill a second gap, the failure of patent law to provide more than 17 years of protection for very non-obvious inventions. An inventor whose invention can be used without revealing it and who believes that it will take much more than 17 years for someone else to independently discover it, has the option of choosing secrecy instead of a patent. He is, in effect, claiming that the lifetime, and thus the social value, of the invention is greater than the patent law assumes—and offering to prove it. The law of trade secrets assists him in this demonstration by increasing the chance that if someone does duplicate the invention, he will do so by inventing it rather than by stealing it.
Trade secret law also provides assistance for the inventor unable to persuade the patent examiner that his invention deserves protection. The inventor offers to demonstrate that his invention is non-obvious by keeping it secret. If someone else invents the same thing next year, that proves the examiner right and no harm has been done. But if the inventor is right and there is no duplication, he gets at least part of the reward he would have gotten if the invention were patentable—as he should, since he has shown that he was right and the patent examiner wrong.
Finally, trade secret law provides protection for an inventor whose invention, while patentable, is not worth the cost of patenting. An invention of only modest value may qualify for a patent but not be worth the substantial cost of getting one. Trade secret protection is cheaper the less valuable the secret, since the less valuable the secret, the less the efforts that will be made to steal it.
In one respect, trade secret law actually provides a more efficient structure of incentives than patent law. Making a patentable invention this year instead of next provides only one extra year's use of the invention—but, under a rule that gives the patent to the first inventor, is rewarded with a seventeen year monopoly. If the cost of making the invention is falling rapidly over time, perhaps due to progress in related technologies, the result will be inefficiently early invention. Under a regime of trade secret protection, on the other hand, the reward to an invention is lower the faster the cost of duplicating it is falling, since an invention that will soon be inexpensive to duplicate will not remain a secret for long. Hence in a regime of trade secret law, unlike one of patent law, an invention whose cost is falling more rapidly will (other things being equal) be made later than one whose cost is falling less rapidly, which is the economically efficient pattern.
And they asked me how I did it, and I gave 'em the Scripture text,
"You keep your light so shining a little in front o' the next!"
They copied all they could follow, but they couldn't copy my mind,
And I left 'em sweating and stealing a year and a half behind.R. Kipling, "The Mary Gloster"
At the beginning of this century, the United States had no copyright treaty with Britain; American publishers were free to pirate books by British authors. Despite the lack of legal protection, British authors routinely collected royalties on American sales. How could that happen?
A British author would provide his American publisher with the manuscript of his book before it was published in England, with the result that the English and American editions came out at about the same time. The pirate publisher could not start setting type until the book was available to be copied. With the printing technology of the time, setting type was a slow and expensive operation, so the legitimate publisher got all of the early sales—which, for most books, meant a large fraction of all sales.
If a pirate edition came out the legitimate publisher could cut his price, issuing a "fighting edition" cheap enough to keep the pirate from making enough money to recover his initial investment. The result of such tactics was that, although piracy occurred, it was not serious enough to keep British authors from making money on their American sales. The publishing technology of the time gave the legitimate publisher enough of a first mover advantage to provide a reasonably good substitute for copyright protection.
A century later, we have almost precisely the opposite situation with regard to computer programs. Having bought a CD with Microsoft Office 98 on it, I discovered that the license does not permit me to use the program on both my home and office machines. As a matter of law, and, in my view, ethics, I am required to buy a second copy of Word in order to be able to work on this book in both places. But if I chose to violate that legal requirement and use the same CD to install the program on both machines, there is nothing Microsoft could do about it.
Computer programs are covered by copyright law, whether the copying is by a seller, a university, or an individual purchaser. But they are, as a practical matter, protected by copyright law only in the first two cases. If someone buys a copy of Office and a CDR drive and proceeds to make hundreds of copies and sell them over the internet, he is quite likely to be caught and prosecuted. If my university decides to cut its expenses by buying one copy of Office and installing it on several hundred machines, a single disgruntled employee who knows how to locate Bill Gates' email address could get us in a lot of trouble. But if I install Office on two machines when I am only licensed for one, or if my five-year-old provides all his friends with copies of his computer games, the chance that either of us will be caught and prosecuted is very near zero.
Software companies have responded to the effective unenforceability of their copyrights in several ways. One early solution was to distribute software on disks designed to be uncopyable by the computers that used them. That turned out to provide enough protection to be inconvenient to legitimate customers but not enough to prevent copying by even moderately sophisticated pirates. Programs appeared designed to copy the protected disks.My favorite episode involved a program for TRS80 computers called "Super Utility Plus" that would (among other things) copy most protected disks but was copy protected against itself. So someone else brought out a program, "SuperDuper," whose sole purpose in life was to make copies of Super Utility Plus.
Eventually, after some unsuccessful litigation by a company that provided software for protecting disks against a company that made software for copying protected disks, most of the software producers gave up and abandoned copy protection. But that was only the first round of the battle; another is being waged even as I write, this time with CD's instead of floppy disks. Currently many programs, especially games, are written to require the presence of the program CD when the program is running. This worked as long as drives for recording CD's were expensive and rare and hard drive space too expensive to be devoted to CD disk images. With hard drives getting steadily cheaper and the price of CDR drives currently near $200 and falling it seems clear, again, that the end is in sight.
If a software firm finds that it cannot protect its intellectual property directly, one solution is to find other ways of making money from it. Those might include selling support, getting revenue from tie-in sales of products that cannot be easily copied, relying on the honesty of its customers, or getting most of its revenue from firms too large to risk being caught with pirated software.
It also might look for new forms of protection. Technologies continue to change, and it is at least possible that changes currently occurring are moving us toward a world where technology will provide a level of protection far beyond what copyright law is supposed to provide.
One important technology is encryption, mathematical procedures for scrambling and unscrambling information. Current developments may make it possible for producers of intellectual property to distribute it in a form that is cryptographically protected, provided with the digital equivalent of a barbed wire fence. The program comes in a digital container; IBM calls their proposed version a crytolope. In order to use the contents, to play a multimedia entertainment package, say, or access a database, you must make a payment, perhaps in digital currency to an online agent of the contents' owner. In exchange for your payment you get controlled access to the contents. Perhaps a dollar buys you an hour viewing of the multimedia or a hundred queries of the database. You can, of course, copy the whole thing, container and contents, and give it to someone else. But if he wants to use it, he too will have to pay.
A variety of projects along these general lines are currently under development; it is still too early to judge whether any will be sufficiently robust to withstand attacks by programmers trying to access the contents without paying for them. One legal issue such technologies raise is whether programs designed to crack such technological protection, digital lockpicks, twenty-first century equivalents of SuperDuper, should themselves be legal. If—as has recently become the case—they are not, and, a much more dubious proposition, if it proves possible to enforce the ban, a container will only have to be strong enough to withstand an attack by an ordinary user, although there is still the risk that an expert might crack the protection and distribute the contents. If programs designed to crack protection are legal, or illegal but widely available, then the protection must be strong enough to withstand a cracking program written by an expert.
At this point the jury is still out. If digital containers turn out to be a viable technology, we will be back in the world of 1900; producers of intellectual property will be able to protect themselves without the assistance of copyright law. And their new protection will be vastly better than the protection they now get from copyright law. Digital containers not only enforce their own restrictions with no need for litigation, they also make possible much more detailed control over the terms on which intellectual property is sold than does copyright.
My Academic Pages
My Home Page
Email to me