Mengmeng Du, MJLST Staffer
Background
In the United States, intellectual property rights in computer software receive protection from copyright law. In 1980, Congress amended 17 U.S.C. § 101 to add software to the subject matters of copyright. Section 101 defines “computer program” as “a set of statements or instructions to be used directly or indirectly in a computer in order to bring about a certain result.” At the same time, Congress added in § 117 exemption of infringement liability under certain circumstances such as when a user installs and runs the software or makes backup copies of the software.
With these seemingly clear definitions, the debate over the extent of the copyrightability of computer software, however, has not abated in the following decades. In Apple Computer, Inc. v. Franklin Computer Corp., 714 F.2d 1240 (3d Cir. 1983), the Third Circuit was asked to determine whether literal copying of computer program object codes constitutes copyright infringement. The Third Circuit ruled that object codes are copyrightable and thus literal copying of such infringes the copyright. In Whelan Ass’n, Inc. v. Jaslow Dental Lab., Inc., 797 F.2d 1222 (3d Cir. 1986) and Computer Ass’n Int’l v. Altai, Inc., 982 F.2d 693 (2d Cir. 1992), the Third and Second Circuit faced the problem of where to draw the line for finding infringement when the copying of software at issue is non-literal. While the Third Circuit would find almost anything below the “purpose of the program” copyrightable, the Second Circuit later developed its more rigorous but more popular “abstraction-filtration-comparison” test, which would yield less copyright protection for non-literal components of computer software.
In 1996, the Supreme Court had a chance to further define the boundary for finding copyright protection in software but missed it. In Lotus Dev. Corp. v. Borland Int’l, Inc., 516 U.S. 233 (1996), the copyrightability of the Lotus menu command hierarchy was questioned. The First Circuit ruled found it an uncopyrightable method of operation by comparing the Lotus menu command hierarchy to the arrangement of buttons on a VCR [see Lotus Dev. Corp. v. Borland Int’l, Inc., 49 F.3d 807 (1st Cir. 1995)]. Lotus petitioned to the U.S. Supreme Court. Due to an even split court with Justice Stevens recusing, the Supreme Court affirmed the First Circuit’s judgment in a per curiam opinion without discussion on the reasoning.
Google v. Oracle
Finally, there is, again, hope to resolve the extent of computer software copyrightability. The Supreme Court granted certiorari to review the decision in Oracle America, Inc., v. Google LLC, 886 F.3d 1179 (Fed. Cir. 2018) on November 15, 2019. In this case, Oracle sued Google for copyright infringement for copying its Java Application Programing Interfaces (APIs) when developing Google’s Android platform. The two parties vehemently debated the copyrightability of the Java APIs and whether the fair use doctrine applies to exempt Google’s use of the declaring code and “structure, sequence, and organization” (SSO) of 37 Java APIs. The Federal Circuit eventually sided with Oracle, finding first in 2014 that the declaring code and SSO of Java APIs are copyrightable (Oracle America, Inc., v. Google LLC, 750 F.3d 1179 (Fed. Cir. 2014)) and then in 2018 that Google’s use is not a fair use (Oracle America, Inc., v. Google LLC, 886 F.3d 1179 (Fed. Cir. 2018)). Google successfully petitioned to the U.S. Supreme Court on its second try. The Supreme Court heard oral arguments from Google, Oracle, and Deputy Solicitor General Malcom Stewart on October 7, 2020.
Professor Thomas F. Cotter
Professor Cotter joined the University of Minnesota Law School faculty in 2006 and is Taft Stettinius & Hollister Professor of Law. With a background in economics and law, Professor Cotter’s principal research interests are in the field of intellectual property law, antitrust, and law and economics. He teaches a variety of intellectual property law courses, including patents, copyright, antitrust, international intellectual property, and patent remedies. For further information, please see his law school profile.
This semester, I attended Professor Cotter’s copyright course, where we studied the Federal Circuit’s decisions in Oracle v. Google. Professor Cotter encouraged the class to listen to the Supreme Court hearing for the now Google v. Oracle case on October 7.
To better understand the law and logic behind Google v. Oracle, I invited Professor Cotter to conduct this blog interview.
The Interview
Q: It is notable that after Federal Circuit’s decision in 2014, Google petitioned to the Supreme Court for the first time but was denied. What do you think is the main reason that the Supreme Court decided to grant cert at this time? Does it have something to do with the “ripeness” in this case, i.e., receiving a final judgement?
A: Like you have suggested, the Supreme Court might have wanted to see what would happen on the fair use issue. Other than that, it is hard to know why the Supreme Court denied cert. It seems like there are a lot of important issues, but often the Supreme Court wants to let them continue to percolate through the lower courts before chiming in, so it can be hard to guess sometimes.
Q: The Supreme Court justices raised a lot of questions during the oral argument. Which one is your favorite question, and why?
A: I’m not sure if I have a favorite question as such, but there were some questions I thought were more getting into the heart of the issue than others.
For example, at pages 80-81 of the transcript, Justice Kavanaugh’s questions to the Deputy Solicitor General Malcom Stewart. These were the two of the more perceptive questions in the entire oral argument. Question number one is on the merger doctrine. Justice Kavanaugh said: “First, Google says in its reply brief that the dispositive undisputed fact in this case is that the declarations could not be written in any other way and still properly respond to the calls used by Java programmers. Are they wrong in saying that?” I think that is a very important question. Justice Kavanaugh then followed that up with a second question on page 81: “And the method of operation, Google says that the declarations are a method of operation because they are for the developers to use, while the implementing code instructs the computer. Your response to that?” I think those are the fundamental questions of the case.
Generally speaking, I would say that I think the better questions were those searching for some kind of analogy, whether it is the QWERTY keyboard or whatever else. But analogies only go so far. Computer software is a thing unto itself. Maybe there is no precise analogy. But you do the best to try to draw inferences from something that is more familiar.
Q: Justice Sotomayor and Oracle disagreed as to whether the precedents have held that there is a distinction between declaring and implementing codes for copyright purpose, whether the precedents have held APIs are not copyrightable, and accordingly, what assumptions the software industry has built on for years. How would you read the precedents?
A: Yes. Particularly precedents from the Ninth Circuit on the question of whether APIs are copyrightable.
I tend to agree with Justice Sotomayor that in these two Ninth Circuit cases in particular— Sega v. Accolade [see Sega Enterprises Ltd. v. Accolade, Inc., 977 F.2d 1510 (9th Cir. 1992)] and Sony v. Connectix [see Sony Computer Entertainment, Inc. v. Connectix Corp., 203 F.3d 596 (9th Cir. 2000)]—those were both cases in which the defendants copied software for the purpose of extracting interface specifications that would enable the defendants to create a compatible program, and Ninth Circuit in both cases held that was a fair use.
The response by Oracle’s counsel was to say among other things that those are fair-use cases and are not going to the fundamental issue of whether the interfaces themselves are copyrightable. But I would say many people have read those cases as premised on the belief that interfaces themselves are not copyrightable and that’s why you can copy the software in its entirety for the purpose of extracting interfaces so as to use them to make compatible programs. So many people would read those cases as standing for the proposition that interfaces are not copyrightable to the extent that they are necessary to enable others to make compatible systems and programs.
So, I’m inclined to think that Justice Sotomayor had the better argument that in construing those cases. But again, they are Ninth Circuit cases and not binding on the decision of the Supreme Court.
Q: So . . . interfaces are not copyrightable is what the industry has understood for years?
A: I think that’s largely true. But I am not an industry insider. There are different opinions depending on who you talk to about whether there is an expectation that someone would pay a license fee to use interfaces, APIs, and the declaring code in particular. There are some instances where companies have paid for that. But my understanding, based on what I have read from the amicus briefs filed in this case, commentaries on it and so on, is that more people are of the view that declaring code was not copyrightable, or at least it was industry custom that you can go ahead and copy it to make a compatible program. Again, not everybody will agree on that, and I am not an insider in the industry. So please take whatever I said with a grain of salt. But based on what I saw, I think that is the dominant view.
Q: I talked to friends in the industry. According to some of them, Google could have developed its own declaring codes or APIs, or paid a “moderate” license fee to Oracle to use the Java SE. But Google chose not to.
A: That’s Oracle’s view, and the view of some commentators and people in the industry.
Here is how I would think of it: there are two viewpoints, and ultimately it comes down to which of the viewpoints the Supreme Court finds more persuasive.
On one hand, Oracle is saying: “You can’t copy our declaring code to make a rival platform. If you want to do that, you would have to ask us and pay us if we can reach an agreement. But you can’t just copy our declaring code to make a rival platform.” This sounds intuitively correct.
But on the other hand, Google comes back and says: “You Oracle cannot use your copyright to inhibit us from creating a rival platform. That would be analogous to Baker v. Selden [see Baker v. Selden, 101 U.S. 99 (1879)], where the attempt was made to leverage copyright to control over an uncopyrightable thing.” So basically, Google is saying that you can’t use your copyright to inhibit others from creating a competing product, as that would be undermining the purpose of copyright and extending copyright to some other endeavors or fields.
In response to that, Oracle says: “But if we can’t assert copyright in our declaring code, the incentive to innovate diminishes. The whole purpose of copyright is to provide that incentive.” I also have long been of the view that many people at least intuitively, rightly or wrongly, feel that if they invest their labor and personality in something, they have some moral entitlement to it, even though you could debate the philosophical issues and how persuasive this really is.
In response to the argument that copyright in declaring code is necessary to validate the incentive to create, Google argues that if the declaring code is copyrightable, then the incentive for people like us to innovate is diminished, because negotiating and paying for the declaring code would give Oracle some control over our creation of the rival platform. This is analogous to the case in Sony v. Universal City Studios [see Sony Corp. of Am. v. Universal City Studios, Inc., 464 U.S. 417 (1984)], where if Sony had lost on the contributory infringement theory, the movie industry would have gained some control over how VCRs and other copying technologies would evolve. Google is also making a point here that were they to develop different declaring codes, it would put their rival platform at a disadvantage since people who are already familiar with the Java declarations would be less likely to use it if they need to learn all these new ones.
That’s where the analogy was made with the QWERTY keyboard. It’s also related to what economists call “network effects”—the value of certain things increases in proportion to the number of other people who are using them. My use of a telephone is negligible if I am the only person in the world who has one. But once more people come onto the network, the value to me of the telephone increases. Similar with the QWERTY keyboard, the network effect provides that if there is one single design in the world, the value of it is much greater and it becomes very difficult for any rival keyboard to ever maintain a position in the marketplace—no one wants to adopt it, even if in theory it is better.
Google would make similar arguments here as well—once people are familiar with the Java declarations, they will be less motivated to learn a new set of declarations to implement the new platform. Therefore, either Google pays for the existing declaration code or makes their own, it diminishes the incentive on Google to develop the rival platform, which enables interoperability for a wide variety of phones and apps.
Thinking beyond this case, if the copyright owners generally have the ability to exert control on declaring codes, maybe that will have the long-term effect of inhibiting innovation and interoperability from which the consumers benefit.
In summary, ultimately, it comes down to which side the Court thinks has the most persuasive arguments.
Q: There is one interesting fact that some people noticed—if you look at how Java originated, Sun actually created Java to break the monopoly of Microsoft. Had Google developed its own declaring codes, it could have ended up with achieving some technology breakthroughs just like Java. Could that be a potential argument to rebut Google’s position regarding inhibition of incentives?
A: Maybe. Network effects are not always insurmountable. Sometimes you might come up with a better product that ultimately does replace the earlier one. Then again, maybe not. People who support Google’s position are concerned that copyright owners having the ability to control the use of declaring code or APIs more generally would ultimately lead to what is called “walled gardens,” which refers to proprietary systems as opposed to open-source systems that enable greater interoperability.
Q: I recall that it was mentioned several times in our copyright law and patents class that it is hard to prove the effects on incentive by evidence. Is it correct to say that is also the case here?
A: Yes, it is. There are a few empirical studies on patent law, and even fewer on copyright law, on this issue.
For example, there may be some empirical evidence showing that the motion pictures industry benefits from having copyright protection. Motion pictures generally take huge amount of money to create. If there is no copyright in motion pictures, it would greatly reduce the incentive to produce, given the high fixed cost and the low marginal cost.
For other works, there is not much empirical evidence one way or the other, either to substantiate that the copyright incentive is necessary or to refute that. Some people would argue that the Oracles of the world would still have very substantial incentive to invest in coming up with new software products. Even if their ability to control the use of some aspects of their software is diminished, there are still substantial benefits to be gained from being the first in the marketplace, e.g., from having good products or from network effects. Maybe the copyright incentive is not altogether necessary. Maybe copyright has more of an inhibiting effect on innovation if it is used too aggressively.
The odd thing about software is that it covers something very functional and the justices were talking about it during the oral argument. It was Congress’s decision, and whether it is a good decision or a bad one, software is copyrightable. Back in the 1970s, there was a debate about whether copyright is a good fit, or maybe it would make sense to have some new and different system in intellectual property law that provides an intellectual property right that lasts for shorter period of time. But the decision was made. Code is copyrightable.
It appears to be some of the justices’ view that the declaring code cannot be viewed as a method of operation because § 101 says code is copyrightable and doesn’t distinguish between declaring code and implementing code. But then you get into a legal doctrine and not the policy. I am not sure whether that argument is necessarily persuasive because it seems you could have a literary work that prima facie looks copyrightable but counts as a method of operation. We will see how the Court resolves this issue.
Q: Justice Gorsuch said it was wise for Google not to linger on the main argument in their brief, i.e., not to make too much Baker v. Selden / § 102(b) arguments. Google did concede that their main argument is the merger doctrine and not the § 102(b) arguments. Do you think it is wise?
A: I am not sure. Some of the justices seem to be skeptical about the Baker v. Selden argument. Though at the end of the day, it seems to me that the idea-expression dichotomy, the merger doctrine, and the Baker v. Selden argument all kind of go to the same issue—all of them refer to § 102(b) which says that you can’t copyright ideas, facts, concepts, systems and methods of operation. From a policy perspective, the idea is that there are certain things are off limits to copyright, and you shouldn’t be able to use your copyright to exert control over those things. So if the majority of justices see this case as implicating that principle, then whether they invoke the merger doctrine, the method of operation principle, or the Baker v. Selden principle, it comes down to the same outcome. But if the majority of justices don’t see this case as so (since Google could have either paid or made its own declaring code), then that analogy is not going to hold.
Q: Several justices have mentioned that other rivals such as Apple and Microsoft didn’t copy to create their competing platform and that Google could have spent the million dollars to develop its own. What do you think about that?
A: That is certainly one way to look at it. The ultimate question is should Google be required to develop its own system that does not require copying the Java declaring code. Maybe that would not be very productive. Allowing programmers to use Java SE may be better for innovation since it is a tool that so many programmers have already known how to use. If Google is to pay for the declaring code or to create its own new ones, there will be a lot of startup costs, which may be socially wasteful. Again, that’s the debate.
Q: Last question. There are many amicus briefs filed in support of Google, but not so many in support of Oracle. Do you think it reflects where the experts stand, and should it substantially impact the Court’s decision (as the Court frequently said that it does not possess the technical expertise to resolve many complex issues)?
A: Amicus briefs may or may not be representative of opinions as a whole. But I think the fact that many more amicus briefs in the case were filed on behalf of Google should at least give some pause. Maybe the amici have a point that code that enables you to make these calls is somehow different from the implementing code. They are all functional in some sense, but declaring code is probably more functional in a general sense and more analogous to a method of operation. This is the way the industry has grown for years. It is the underlying assumption of many people in the industry that it is perfectly lawful to do this. Maybe the Supreme Court should at least give serious consideration whether it should run up against the custom, since many people in the field of computer science and as well law are of the view that Google’s argument is more sensible. But again, there are people who disagree with that, and the Supreme Court has to evaluate all of the opinions.
(the end of the interview)
Closing
As Professor Cotter has pointed out, the debate behind Google v. Oracle comes down to the core issue of why we should provide copyright protection for computer software. Each side has important interests at stake—Oracle’s interest in guarding its investment of labor and personality in Java and Google’s interest in being free from inhibition of innovation. Society at large also has an interest in having a balanced intellectual property system that provides most incentive for people to create.
The 83 computer scientists mentioned in the amicus briefs are of the point that the sky will fall if the Supreme Court rule against Google in this case. Whether it is true or not, this time, the future of the software industry is really at stake. All we can do is wait and see what the Supreme Court will say about these important issues in months.