SKELJ's Comprehensive Oracle v. Google Writeup - First Published 05/23/2021 - Last Update: 05/23/2021
An "absurd ruling."
A "major victory for . . . innovation."

This is a walkthrough (intended for desktop browsers) of the April 2021 Supreme Court of the United States (SCOTUS) decision Oracle v. Google (OvG) and its lower court decisions. Like layers of an onion, you can read just the TL;DRs and overviews for a quick understanding or descend into the deep nitty-gritty of the opinion writeups and precedential case summary constellation. The intended audience is mostly developers and non-developers rather than attorneys (though attorneys are welcome, you'll probably just want to skip the large background portions).

Originally, I was just going to post my handwritten notes, but several motivations precipitated this more lengthy writeup, e.g.:

  • Several developers specifically asked for an informal writeup like this, some wanting a quick summary, some wanting a comprehensive overview
  • Misunderstandings of OvG are rampant with real-world consequences for software ecosystems
  • Hyperbolic news articles, confusing commentators, and self-serving statements by agenda-motivated amici and "legal intellectuals" have complicated an already challenging fact pattern and precedential landscape
  • In 1995, Judge Boudin warned that a decision like this would result in judicial inefficiency. The best way to combat that inefficiency is through education. Particularly, litigation follows from divergent expectations between parties - the more clarity in the legal landscape, the less chance of such litigation inducing disparate expectations

Before jumping in: DISCLAIMER: This is *not* legal advice. My language is deliberately loose and non-legal so as to convey general concepts rather than legally precise propositions - there are almost certainly typos (I've done this write up ad hoc as I find time outside work hours; it's not perfect, but I need to move on to other obligations). Also, as will quickly become evident, it's impossible to apply anything herein without having an attorney carefully consider a specific set of facts - so don't. I'll also probably tweak and update portions of the writeup as I receive feedback. So use this to guide your understanding and your own further reading, *NOT* for application to your or anyone else's current situation. If you have a legal question, consult with an attorney.

How to use this write up:

  • Click on any section in the Table of Contents (TOC) on the left side of the page to jump to that topic. DO bounce around. The TOC will highlight where you are if you get lost. This is meant more for review on a desktop rather than on a mobile phone.
  • If you only have 5 seconds to live, read the one line TL;DR summary.
  • If you just want the brutally quick, 5-minute summary, read the very non-legal / non-technical TL;DR. If you want a deep dive, you may still find such a high-level summary a useful foundational context.
  • If you want a slightly more legally substantive summary, read the more legally substantive TL;DR.
  • Before jumping into the OvG walkthroughs, I've previously noticed when describing OvG that it's best not to assume my audience's background. Accordingly, you might consider first reviewing an item from the "Various General Background", e.g.:
  • The case walkthroughs go through the opinions item by item.
  • If you want a REALLY deep, deep dive, I also threw my informal preceding case notes into an interactive table.
  • There's a FAQ. If you have more questions, shoot me an email.

SCOTUS has accepted the invitation posed by Judge Boudin in 1995 to NOT treat software APIs the same as dramatic literary works under fair use, at least insofar as the API owner actively encouraged third parties to use the API as a locus for their own creative development prior to asserting infringement against the API copyist (in contrast to, say, characters in a novel which were merely made familiar to readers, but which the author did not encourage readers to use in the generation of their own works).

(N.b. that is a very paraphrased summary - the decision is considerably more nuanced)
  • Congress creates laws, the Supreme Court interprets them. Interpretation can itself, in effect, expand or contract the law.
  • The Copyright Act is a law. It includes three parts:
    1. A part indicating what can be copyrighted.
    2. A part indicating what rights copyright confers to the author.
    3. A part ("fair use") limiting the rights in #2 as to certain activities by potential infringers. Congress intentionally left this portion open for courts to balance on a case-by-case basis.
  • Programmers control computers by writing in a pseudo-language called code. For convenience, code may be run by calling "shorthand names" instead of copying large blocks of code to which those names refer. The collection of shorthand names are known as an Application Programming Interface (API).
  • Google wanted to avail itself of developers' familiarity with the JAVA API when creating its Android smart phone. It copied the JAVA API, but not the underlying blocks of code it referenced. Google rewrote those blocks themselves for the Android phone. Oracle (the successor owner of JAVA) sued Google for its copying of the API.
  • APIs represent a nebulous portion of the Copyright Act, filled with internal tensions.
    • On the one hand, the Act explicitly calls out "compilations" as being deserving of protection in #1 above. Software code is likewise explicitly protected in #1. #2 also explicitly protects "derivative works" of that code.
    • However, software is also afforded less protection under #1 than purely fanciful, non-functional works, like novels. #3 also affords protection for "transformative" use of copyrighted material, despite this possibly being at odds with derivative works, which are a right protected under #2.
  • The nebulousness is further complicated by the fact that #3 is a heavily fact-dependent analysis. Consequently, very similar situations may result in radically different outcomes.
    • This complication has been recognized since the 1995 case Lotus v. Borland, also dealing with a software interface, wherein Judge Boudin explained that fact patterns with mostly non-fanciful APIs confront a "Scylla and Charybdis" in their resolution:
      • Scylla: Denying copyright to an interface under #1 is simple and introduces less ambiguity, but is perhaps less consistent with the Act, which affords copyrightability to even small amounts of expression, even if they also fulfill a functional purpose
      • Charybdis: Acknowledging copyrightability under #1, but applying fair use under #3, may be more consistent with an Act, but as #3 is heavily fact dependent, this risks inefficiency and uncertainty. While such inefficiency may create more work for attorneys, it's likely not very good for software businesses or for developers.
  • In a first thread of appeals addressing copyrightability of the API, the Court of Appeals for the Federal Circuit reaffirms that APIs may be copyrighted under #1. The Supreme Court does not grant cert to contest this conclusion.
  • In a second thread of appeals addressing Fair Use of the API, the Supreme Court majority determines, under these facts, that Google's usage is a Fair Use under #3 (effectively choosing Judge Boudin's "Charybdis").
    • The Supreme Court dissent does not consider Google's behavior sufficient for fair use. The disagreement between the Supreme Court majority / dissent itself demonstrates dangers inherent in the "Charybdis" approach of which Judge Boudin warned
  • Going forward, consonant with the "Charybdis" approach, analyses will be very fact intensive and complicated by the heterogenous fair use case law (discussed below). It is possible, but unlikely, that Congress may amend the Copyright Act to provide more clarity.

The CAFC and Supreme Court decisions basically result in the following four "takeaways" affecting the current legal landscape:

  1. The Seventh Amendment does not guarantee that a jury must decide fair use - rather, where a jury is employed, it is only for factual determinations and reviewed per application of the recent U.S. Bank Supreme Court case (hopefully, this will mitigate decision variability)
  2. API taxonomies may be copyrightable - so long as they satisfy the various requirements under the Copyright Act even if their expressive lines serve a functional purpose.
  3. As per Campbell v. Acuff-Rose Music, the fair use factors should be holistically considered (at least for software) - categorical precedence of some factors, such as market effect, would not seem to be viable post-OvG. I emphasize "seem" because the court explicitly says they aren't overruling precedent, but that may be the effect in some fact patterns.
  4. Software Fair Use is NOT the same as Literary Fair Use - The court basically responds in the affirmative to the two queries posed in Lotus v. Borland by Judge Boudin's concurrence.

Regarding that last point, Judge Boudin absolutely nailed the issues here back in 1995 in his concurrence (provided below) to Lotus v. Borland, basically, making two critical observations:

  1. It's not clear if Congress, in the Copyright Act, wanted courts to treat software exactly like literary content, or as something "different" with its own corpus of analysis for copyrightability, fair use, etc.
  2. If software is "different" from literature under the Act (e.g., developer investment in an API is NOT the same as readers familiarizing themselves with characters in a novel), then the basis upon which a party prevails is tricky and significant. Saying that certain software is not protected (what they did in Lotus) may not be as proper as employing a "privileged use" / fair use doctrine, since the latter would distinguish original author-creator investment (which we want to protect/reward) from user/developer investment (which we also want to encourage and so should lessen copyright protection when it applies). (I am seriously paraphrasing Judge Boudin here - you should go read the original)

OvG is basically providing answers to these questions. Saying:

  1. Yes, we're going to treat software differently.
  2. Yes, we're going to use fair use to recognize and respect that distinction.

Going Forward:

  • Copyright Act Tension / Judicial Inefficiency -

    There is considerable tension between the SCOTUS decision and the Copyright Act. Judge Boudin predicted this, cautioning against blithely choosing this fair use / privileged use approach.

    "But a privileged use doctrine would certainly involve problems of its own. It might more closely tailor the limits on copyright protection to the reasons for limiting that protection; but it would entail a host of administrative problems that would cause cost and delay, and would also reduce the ability of the industry to predict outcomes. Indeed, to the extent that Lotus' menu is an important standard in the industry, it might be argued that any use ought to be deemed privileged."
    Judge Boudin's concurrence, Lotus v. Borland (emphasis added)

    Having "any use" "deemed privileged" for ALL APIs would effectively eviscerate protection for APIs under the Copyright Act, which probably isn't what Congress intended. Thus, unless Congress steps in, courts are going to want to resolve that tension, finding some "true negative" datapoint that usage of an API wasn't fair use so as to counterbalance this true positive, thereby avoiding an implication that OvG eviscerates protection of all API structures under the Act.

    The SCOTUS majority are smart folks and are aware of this. They all but make Congress an explicit, handwritten invitation to step in and clarify 17 USC 107:

    "The fact that computer programs are primarily functional makes it difficult to apply traditional copyright concepts in that technological world. See Lotus Development Corp., 49 F. 3d, at 820 (Boudin, J., concurring). In doing so here, we have not changed the nature of those concepts. We do not overturn or modify our earlier cases involving fair use—cases, for example, that involve “knockoff ” products, journalistic writings, and parodies. Rather, we here recognize that application of a copyright doctrine such as fair use has long proved a cooperative effort of Legislatures and courts, and that Congress, in our view, intended that it so continue. As such, we have looked to the principles set forth in the fair use statute, §107, and set forth in our earlier cases, and applied them to this different kind of copyrighted work."
    SCOTUS Majority, OvG (emphasis added)

    Anyway, as I said, courts will be on the hunt now for a "true negative, i.e. not fair use" case to demonstrate that fair use hasn't eviscerated the Act. There may be several such cases and SCOTUS may not grant cert on all (or any) of them - so don't be the one to blithely rely upon OvG for your situation - indeed, even if they did grant cert, it may only be so that your case can serve as the happily discovered "sacrificial" complement for SCOTUS to demonstrate that their Fair Use analysis didn't eviscerate the Act.

  • Open Source Balkanization -

    The SCOTUS decision, implicitly, provides guidance for avoiding open source ecosystem balkanization. It's not very definitive or reassuring guidance, but it is there. Given that shaky ground, trademark and patent should be rolled in (to the extent possible) if you're managing an ecosystem and really want to strengthen your position (appreciating that there is precedent, though, guiding what you can and can't do in this regard - see, e.g., trademark and the Sega decision, as well as the recent Alice patent decisions). People often overlook, e.g., the patent language in the Apache license - it might be time to brush it off and tweak it to fill any lacunae for interface control resulting from OvG in your situation.

Some developers have confessed to me that they see legal proceedings as mere "theater", lacking rigor, and governed mostly by the whims of judges and lawyers.

This is understandable.

Most schools no longer teach dialectics, forensic rhetoric, and the trivium progression from grammar->logic->rhetoric, etc., at least not explicitly, in the class room.

Indeed, most law schools do not.

Rather, law schools tend to be so single-mindedly focused upon training their students to pass the bar [presumably because their US ranking depends in large part upon those numbers] that they rarely, if ever, discuss legal philosophy and the underpinnings of forensic rhetoric and analysis. I once had a lovely discussion regarding Kotaro Tanaka with a Political Philosophy Professor at a prestigious university here in the States. I remarked that their law school was lucky to have him. He chuckled saying that he had never been invited to speak at the law school and he never expected to be.

Instead, many lawyers pick up these tools implicitly during their apprenticeship. Consequently, they aren't able to articulate that they're, say, applying doctrine from a 17th Century Dutch humanist or a syllogism with term distributions in the categorical propositions of one of the A, E, I, or O forms - they just do so, not knowing that that's what they're doing (in fairness, most of the top judges and litigators I've met, do know, but they're the rare breed).

In any event, the result is that many engineers I know view linguistic reasoning with suspicion, certainly as compared with the more hallowed ground of mathematics.

If you've experienced this skepticism, I invite you to suspend judgment for a moment and consider the following.

Words, like numbers, have meaning.

As in math, you can reason with words. As in math, you can apply logic with words. As in math, you can state propositions and deduce conclusions with words.

A skeptical developer's objections are typically as follows:


  • Objection #1: Words, unlike numbers, are vague. When I say "cat" does the definition include an animal with stripes? A tail? I can find cats without tails. Heck, I can find cats without stripes. Now the number 0? Multiplication? Exponentiation? We all know what those are.
  • Objection #2: Following from objection #1, math is "objective", its concepts have "inherent meaning" and can therefore be used for reasoning. Words, in contrast, are subjective. Their concepts lack inherent meaning. Therefore they cannot be used for reasoning to arrive at objective conclusions.

To which I respond as follows:

  • Response #1: All definitions, whether mathematic or linguistic, require context to have meaning. This is known as "coming to terms", i.e., before having a discussion we're going to agree on our definitions. What's 00? Take the limit from 0+ and it's positive 1. Take the limit from 0- and it's negative 1. Take the algebraic definition of exponentiation where, e.g., a2 is 1 x a x a, and 00 is 0. Seek a continuous limit across the reals and it's undefined. Context dictates meaning. It's not a coincidence that foundational propositions in both grammar and mathematical logic are referred to as sentences.
  • Response #2: Even when context has been established, definitions still suffer from boundaries between "difference in degree" and "difference in kind." What's the boundary for statistical significance? 0.05? Why not 0.050001? This is well recognized in both mathematics and linguistics - sorites paradox. So where necessary, we must address the boundaries of our definitions if we are to reason well.

In any event, in law, we're going to reason with words.

Per response #1 above, reasoning never occurs in vacuum. There are always foundational propositions, e.g.

  1. You are human.
  2. You agree that reason/logic are viable methods for making deductions.
  3. Etc.

In law, in addition to those basic foundational propositions, we are ALSO required to include certain law-specific propositions, arising from:

  • Statutes
  • Previous cases
  • Facts (as recognized by the court, and yes, this can be a contentious process)

Pet peeve: note that I didn't say amicus briefs, law review articles, or your mom's blog - there are good philosophical reasons why we don't just look anywhere to justify our legal reasoning. Different entities in the legal system have different functions. Congress, as the "voice" of the people, makes the laws and judges interpret those laws. Naturally, interpreting (statutory or common) law affects the effect of those laws, and this is the basis for the "Previous cases" item above. Congress may say "All cats need leashes" and the courts may then say "cats are understood to have stripes, ocelots don't have stripes, so ocelots don't need leashes." If the courts get too creative like this, Congress will rein them back in by amending the law (this was what happened in the 1992 amendment to the Copyright Act when Congress didn't like how the courts were considering nonpublication in their fair use analysis). Until Congress does that, though, lower courts are bound by (i.e., must incorporate propositions from in their reasoning) the interpretations of the higher courts.

This is not a perfect system, but it is a good system. I put "voice" in air quotes above, because lobbyists and elected officials obviously impose some distance from the average American and Congress. But that's better than having judges make law ab initio. Lobbyists, private interests, etc. spend months, years, analyzing their positions. While they may not always have the public weal in the forefront of their mind, they tend to really know the issue. A judge, in contrast, is handling dozens of cases at one time, has, at best, usually only a few months (typically much less) to consider an issue, may not even be elected by the people, and is relying upon relatively short briefs to inform their analysis. Consequently, judges should, and often do, correctly perceive their role as one of only modest interpretation. They appreciate that even if they have the best of intentions, they just aren't in a position to usurp the legislative process.

Let's discuss in more detail with a non-legal hypothetical that actually explains some of the behavior in OvG.

Imagine (or recall) that you have a spouse.

You live on the FAR WEST side of a small town. You are running errands at the hardware store on the FAR EAST side of town. You receive a text from your spouse, which reads:

"Hey babe - on the way home, can you stop by my favorite restaurant and pick up dinner? Luv ya"

You break into a feverish sweat, for it is a culinary curiosity of your city of residence that there are only two restaurants your spouse frequents, on opposite sides of town: Restaurant A in the north and Restaurant B in the south.

Thus, your situation is as follows:

Travelling to both is unreasonable. Without texting back and admitting ignorance as to your beloved's preferences, how will you choose?

What follows is basically forensic rhetorical (i.e., legal) reasoning. It's as if you have a demon for Restaurant A on one shoulder and a demon for Restaurant B on the other (no angels, this is law after all) advancing arguments based on your experience, e.g.:

  • Demon A:
    • "Obviously they're referring to A. Isn't that where they asked to have their birthday party? Why suggest that if it wasn't their favorite?"
  • Demon B:
    • "No, no, no. It's obviously B. That's where you proposed for heaven's sake. What spouse would call the other place their favorite after that?"
  • Demon A:
    • "I'm telling you, it's A. You know they love Italian."
  • Demon B:
    • "You know they're allergic to garlic bread."
Etc., etc. you get the idea. The existing evidentiary / proposition base is being harvested for arguments. For now, assume both sides' arguments are equally consistent with the available propositions.

I refer to this space of viable reasoning from the available, agreed propositions, the "Fog of War". The larger the Fog of War, the more perspectives are available for making viable arguments, but the less certain, and consequently less efficient in reaching a conclusion, is the reasoning landscape.

(Arguably, the only difference between mathematic and linguistic reasoning is that the highly artificial nature of the former facilitates stronger Fog of War control)

For example - let's continue the hypothetical and narrow the landscape by adding a proposition. Your spouse now sends a second, follow up text, which reads:

"Oh, and would you also pick up some milk on your way home?"
There is only one grocery in town, situated as follows:

Has this dispositively resolved the question in favor of Restaurant A?

No. It's possible your spouse intends you to make a very long trip home, going to Restaurant B.

But the Fog of War has been so contracted (this is a variation of Occam's razor) such that there are fewer and less viable arguments now available to Demon B.

To be clear: Demon B's arguments are not now logically incorrect - they begin from viable propositions and reason validly with logical rules. They are, however, less consistent with the proposition base than are the arguments of Demon A, since Demon B must exclude, ignore, or otherwise dispense with, more propositions than Demon A to reason validly (if your mind has conjured an image of a dot product, yeah, that's basically the idea, though it's more nuanced for complicated arguments; and yes, what age would we be living in if there weren't folks trying to automate all this).

Congress and the courts are aware of this interplay between argument flexibility and propositional contraction and they, just as your spouse, may encourage one or the other deliberately.Example

For example, let's say your wise and loving spouse wants to express their preference, but does not want you to feel guilty if you choose incorrectly. To that end, they may deliberately withhold sending that follow up text message. This way, hopefully you bring home what they prefer, but if not, they can still smile and maintain positive conjugal relations. These benefits, however, come at a cost: your spouse may not get the meal they desired, you may waste time and energy pining and trying to guess their intentions, etc. If your spouse prioritizes efficiency (e.g., if you've screwed up enough times), they may send the follow up text or state their preference outright.

Congress and the courts do the same thing. We live in a messy, changing world. Sometimes we want courts to reason on a case-by-case basis, accommodating broad Fogs of War as circumstances change ("What are the bounds of free speech?"), and sometimes we want them to decide quickly and move on ("No, you can't dump toxic waste on city hall, period, no exceptions."). Usually businesses and Wall Street prefer the latter, narrower Fogs of War, since it provides certainty in their own analyses. Typically, that means strong, clear property rights - good fences make good neighbors. However, there are some business models that benefit from broad Fogs of War, specifically preferring ambiguity around property rights.

Anyway, the point here is that we don't read court opinions, or each side's arguments, in a vacuum. We read each word in each sentence, within the larger propositional context. Arguments less consistent with the propositional context are as objectively less preferable as mathematical proofs that violate or misuse axioms (which shouldn't be surprising, since the latter are [arguably] specific instances of the former). Of course, with a broad enough Fog of War, two arguments may be equally consistent, even if they come to opposite conclusions (as with your initial decision whether to select restaurant A or B).

As a final thought, note that while new precedential propositions usually narrow the Fog of War, new precedential propositions may sometimes increase the Fog of War. That is, the Fog of War usually enlarges due to "deductive ambiguity" ("How does this specific instance align with a general principle?"), but it may also enlarge from "inductive ambiguity" ("Having seen this specific instance, what is the corresponding general principle?").

For example, if a statute says "cats must wear leashes" (general principle) that's pretty unambiguous. Initially, that may be an easy rule to follow. However, if a judge says "that doesn't apply to ocelots in the zoo" (specific instance) then you have to reconcile the two, and the once narrow Fog of War is now much broader. What was it about ocelots in zoos that the judge thought was excluded from "cats"? The more explicit the judge is in their reasoning, the less risk there is for inductive ambiguity ("In the Congressional hearings for the statute, Congress explained that they were concerned about violent incidents with home-owned predatory felines, not those in the controlled environments of zoos - so I'm excluding zoo-managed ocelots . . ."). If the judge wasn't clear, though, about their reasoning, this conclusion may have taken an otherwise clear definition ("cats") with a limited Fog of War and modified it to something unclear with a consequently larger Fog of War ("cats, but somehow not ocelots affiliated with zoos . . . somehow . . . maybe because they don't have stripes? They're behind bars? . . .").

This is why, in terms of precedential effect, a judge's REASONING can often be more important than the judge's CONCLUSION. Reaching the "right" conclusion by vague reasoning can be, in many ways, worse than reaching the "wrong" conclusion by clear reasoning, since the latter reveals the judge's interpretive process, whereas the former resembles usurpation of Congress' legislative function. Demon B may luck out after you receive the second text, as your spouse may have actually intended Restaurant B - but that certainly wasn't as consistent with the propositional landscape as posed in the hypothetical as was the alternative.

Inductive ambiguity is predominant in the "Charybdis" of Judge Boudin's warning discussed above.

Rhetoric and dialectic are fascinating subjects (not just as applied to legal reasoning) - this is just a glimpse to inform various of my observations on OvG in this writeup. If you're interested in learning more, reach out and I can recommend some books / articles.

Courts don't let you sue someone for anything. Having a bad haircut is not an actionable offense (whew). Only certain "rights" grounded in legal propositions are recognized.

For example, a contract gives you rights with respect to a specific entity - the counterparty to the contract. In contrast, "property" is a set of rights as to everyone in the whole world (well, everyone that's within the jurisdiction of the court). E.g., I don't need to know someone personally or have a contractual relationship with them to be liable for trespass upon their property. Hanging out on their lawn is trespass even if they never saw me before.

The foundations (in the United States) for the copyright property rights are found in the Patent and Copyright Clause of the Constitution, which provides Congress the power "[t]o 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."

Congress has instantiated that power with the "Copyright Act" (appearing in 17 U.S.C. §§ 101-810 of the US code) whose last major revision was in 1976 (though there have been minor adjustments since). The 1976 Amendment addressed both advances in technology and the United States' imminent participation in the Berne Convention, an international agreement on copyright.

As most folks are generally aware, copyright protects expressive works.

"Copyright protection subsists, in accordance with this title, in original works of authorship fixed in any tangible medium of expression . . . Works of authorship include the following categories: (1) literary works . . . "
(As highlighted above, note that the work must be "fixed." A message riding on an RF carrier wave doesn't receive a copyright. In contrast, once the message is saved to your hard drive, it's (typically) been "fixed.")

Just like, say, a cookbook, a computer program is a "literary work" and therefore receives a copyright.

" . . . 'Literary works' are works, other than audiovisual works, expressed in words, numbers, or other verbal or numerical symbols or indicia, regardless of the nature of the material objects, such as books, periodicals, manuscripts, phonorecords, film, tapes, disks, or cards, in which they are embodied."
Now, even though that doesn't say "computer programs" specifically, the legislative history was pretty clear:
The Copyright Act affords protection to "original works of authorship fixed in any tangible medium of expression . . . ." 17 U.S.C. § 102(a). This broad category of protected "works" includes "literary works," id. at § 102(a)(1), which are defined by the Act as works, other than audiovisual works, expressed in words, numbers, or other verbal or numerical symbols or indicia, regardless of the nature of the material objects, such as books, periodicals, manuscripts, phonorecords, film tapes, disks, or cards, in which they are embodied. 17 U.S.C. § 101. While computer programs are not specifically listed as part of the above statutory definition, the legislative history leaves no doubt that Congress intended them to be considered literary works. See H.R.Rep. No. 1476, 94th Cong., 2d Sess. 54, reprinted in 1976 U.S.C.C.A.N. 5659, 5667 (hereinafter "House Report"); Whelan, 797 F.2d at 1234; Apple Computer, 714 F.2d at 1247.
Computer Associates International, Inc. v. Altai, Inc., 982 F.2d 693 (2d Cir. 1992)
Courts have also since confirmed that BOTH source and object code are protected.
"It is well, though recently, established that copyright protection extends to a program's source and object codes. Stern Elecs., Inc. v. Kaufman, 669 F.2d 852, 855 n.3 (2d Cir. 1982) (source code); Apple Computer, Inc. v. Franklin Computer Corp., 714 F.2d 1240, 1246-47 (3d Cir. 1983) (source and object code), cert. dismissed, 464 U.S. 1033, 104 S. Ct. 690, 79 L. Ed. 2d 158 (1984); Williams Elecs., Inc. v. Artic International, Inc. 685 F.2d 870 (3d Cir. 1982) (object code)."
Whelan Associates, Inc. v. Jaslow Dental Laboratory, Inc., 797 F.2d 1222 (3d Cir. 1986)
In 17 USC § 106 the Act confers upon the owner of the literary work the following exclusionary rights :
  1. "to reproduce the copyrighted work in copies"
  2. "to prepare derivative works based upon the copyrighted works"
  3. "to distribute copies or phonorecords of the copyrighted work"
  4. "to perform the copyrighted work publicly"
  5. "to display the copyrighted work publicly"
Again, note that copyright extends to the expressive character of the literary work. If a cookbook has a recipe on cheesecake, these rights apply to the expressive presentation of that recipe - you can't stop someone from making the cheesecake itself. Neither can you stop someone from describing the recipe in a different expressive character (so long as that expressive character isn't a mere derivative of the original). Ditto code. While the "expressive" aspects of software may receive a copyright, the "ideas" embodied therein, do not.

This is known as the "idea-expression dichotomy" and is made explicit in the Act:

"(b) In no case does copyright protection for an original work of authorship extend to any idea, procedure, process, system, method of operation, concept, principle, or discovery, regardless of the form in which it is described, explained, illustrated, or embodied in such work."

(Some developers guffaw at this: "You mean my lean, minimalist, functional code is less amenable to copyright protection than some bonehead's circuitous, gobblygook mess?"


But typically even a lean approach will receive some protection.)

Courts have clarified that the "expressive" portion of the literary work must be "original"

"Originality is a constitutional requirement. The source of Congress' power to enact copyright laws is Article I, § 8, cl. 8, of the Constitution, which authorizes Congress to "secure for limited Times to Authors . . . the exclusive Right to their respective Writings." In two decisions from the late 19th century -- The Trade-Mark Cases, 100 U.S. 82 (1879); and Burrow-Giles Lithographic Co. v. Sarony, 111 U.S. 53 (1884) -- this Court defined the crucial terms "authors" and "writings." In so doing, the Court made it unmistakably clear that these terms presuppose a degree of originality.
. . .
In The Trade-Mark Cases, the Court addressed the constitutional scope of "writings." For a particular work to be classified "under the head of writings of authors," the Court determined, "originality is required." 100 U.S., at 94. The Court explained that originality requires independent creation plus a modicum of creativity: "While the word writings may be liberally construed, as it has been, to include original designs for engraving, prints, &c., it is only such as are original, and are founded in the creative powers of the mind. The writings which are to be protected are the fruits of intellectual labor, embodied in the form of books, prints, engravings, and the like." Ibid. (emphasis in original). . . . In Burrow-Giles, the Court distilled the same requirement from the Constitution's use of the word "authors." The Court defined "author," in a constitutional sense, to mean "he to whom anything owes its origin; originator; maker." 111 U.S., at 58 (internal quotation marks omitted). As in The Trade-Mark Cases, the Court emphasized the creative component of originality. It described copyright as being limited to "original intellectual conceptions of the author," 111 U.S., at 58, and stressed the importance of requiring an author who accuses another of infringement to prove "the existence of those facts of originality, of intellectual production, of thought, and conception." Id., at 59-60. "
Feist Publications, Inc. v. Rural Telephone Services Co., Inc., 499 U.S. 340, 111 S. Ct. 1282, 113 L. Ed. 2d 358 (1991)
You need to show that the defendant copied that original, expressive portion to prevail:
"Once the plaintiff has shown that it holds a valid copyright, it must next prove that the defendant unlawfully appropriated protected portions of the copyrighted work. This question involves two separate inquiries: 1) whether the defendant, as a factual matter, copied portions of the plaintiff's program; and 2) whether, as a mixed issue of fact and law, those elements of the program that have been copied are protected expression and of such importance to the copied work that the appropriation is actionable. 3 MelvilleB. Nimmer, Nimmer on Copyright 13.01[B], at 13-8 to 13-15 (1993) (hereinafter "Nimmer"); see Arnstein v. Porter, 154 F.2d 464, 472-73 (2d Cir. 1946)."
Gates Rubber Co. v. Bando Chemical Industries, Ltd., 9 F.3d 823 (10th Cir. 1993)

Finally, and importantly for our discussion of APIs, note that the Act also explicitly grants protection to compilations:

"(a) The subject matter of copyright as specified by section 102 includes compilations and derivative works, but protection for a work employing preexisting material in which copyright subsists does not extend to any part of the work in which such material has been used unlawfully.
(b) The copyright in a compilation or derivative work extends only to the material contributed by the author of such work, as distinguished from the preexisting material employed in the work, and does not imply any exclusive right in the preexisting material. The copyright in such work is independent of, and does not affect or enlarge the scope, duration, ownership, or subsistence of, any copyright protection in the preexisting material. "

In summary, you should take away the following points:

  1. Software, to the extent it is a fixed "original" "expression", receives a copyright
  2. Copyright provides the owner with certain exclusionary rights
  3. However, the Idea-Expression Dichotomy limits those rights to only the "original" "expressive" portions of the code

That's it. As far as IP regimes go, we could have stopped there. However, the Act also limits copyright enforcement based upon a doctrine we'll discuss in the next section: fair use.

Fair Use was originally a judicially created doctrine stretching back to the 1700s (e.g., in 1740 the English Court of Chancery distinguished shortened copies of works, which were "fair abridgements", from infringing copies). This is more than just a fun fact, as the manner in which such doctrines were applied in 1791 determines, e.g., whether the 7th amendment guarantees a jury trial for fair use today.quote

Why have fair use? Why not just have absolute copyright monopolies? Part of the philosophical underpinning here arises from the Constitution's command to "promote the Progress of Science and useful Arts", not "reward authors / publishers indiscriminately."

"The ultimate aim is, by this incentive, to stimulate artistic creativity for the general public good." Twentieth Century Music Corp. v. Aiken, 422 U.S. 151, 156, 45 L. Ed. 2d 84, 95 S. Ct. 2040 (1975). Thus, the copyright law seeks to establish a delicate equilibrium. On the one hand, it affords protection to authors as an incentive to create, and, on the other, it must appropriately limit the extent of that protection so as to avoid the effects of monopolistic stagnation. In applying the federal act to new types of cases, courts must always keep this symmetry in mind. Id."
Computer Associates International, Inc. v. Altai, Inc., 982 F.2d 693 (2d Cir. 1992)

As the doctrine solidified in the United States, it was ultimately codified in 17 USC § 107

"Notwithstanding the provisions of sections 106 and 106A, the fair use of a copyrighted work, including such use by reproduction in copies or phonorecords or by any other means specified by that section, for purposes such as criticism, comment, news reporting, teaching (including multiple copies for classroom use), scholarship, or research, is not an infringement of copyright. In determining whether the use made of a work in any particular case is a fair use the factors to be considered shall include
  1. the purpose and character of the use, including whether such use is of a commercial nature or is for nonprofit educational purposes;
  2. the nature of the copyrighted work;
  3. the amount and substantiality of the portion used in relation to the copyrighted work as a whole; and
  4. the effect of the use upon the potential market for or value of the copyrighted work.

"Shall" and "include" are magic legal words. "Shall" means "you have to." And "include" means, "these things, but not just these things, there may be more . . . "

The "Purpose", "Nature", "Amount", and "Effect" factors are referred to as the "four factor" test of fair use. I usually abbreviate them as the PNAE factors.

For convenience, you'll see me put a court's analysis in a table like the below:

I'll discuss the Purpose factor here.
I'll discuss the Nature factor here.
I'll discuss the Amount factor here.
I'll discuss the Effect factor here.

So far so good.

Now the hiccup here is that to apply this test, we have to balance those factors in view of:

  1. A given set of facts which can (usually) be creatively interpreted in a wide variety of ways; and
  2. A voluminous case law, which can itself (usually) be interpreted in a wide variety of ways (I've summarized some of the more relevant cases here).

What's more, the Act, not just this fair use portion, but as a whole, is a "one-size-fits-all" document that's being applied to purely literary contexts (e.g., novels) as well as to much more technical domains (e.g., power plant software).

As you can imagine, this creates a HUGE Fog of War, leaving some exasperated judges to declare that the fair use doctrine is:

"[T]he most troublesome in the whole law of copyright."
Dellar v. Samuel Goldwyn, Inc., 104 F.2d 661, 662 (2d Cir. 1939)

Ideally, courts would gradually clarify the doctrine's application over time, slowly annealing it into an efficient, analytic steady-state. Alas, there's this thing that keeps throwing a wrench in the machine: technology.

If you're a developer, skip this section.

Computers understand 1s and 0s, humans understand words and phrases.

Programmers tell a computer what to do by writing “words and phrases”, i.e., code, which is then converted to the corresponding 1s and 0s (typically by a thing called a "compiler").

For the programmer’s convenience, “words and phrases” can be grouped into functions, just like math functions.

Let's say, e.g., that a developer wants to multiply a number by 1000 add two and divide by 3, for each of the numbers 2, 3, 4. They could say:


but that gets bloated and tedious pretty quickly. If, instead, they define a function

f(x)=(x*1000+2 )/3

They can just say:


That's a lot less typing and also makes code easier to read.

Now, real software functions are considerably more complicated, but they still follow the same principle. Here's a more realistic example.

In C++, the line of code

print(‘hello world’);

tells the computer:

“Dear computer. Please go find the function called ‘print’ and give it the phrase ‘hello world’.”

A computer is like a lobotomized cheetah: very fast and very stupid.

Milliseconds after receiving this code, the computer may say:

“No. Fail. Not going to do it. What the heck is ‘print’? I don’t know where to find any ‘print’, let alone give it the input 'hello world.'”

With a heavy sigh, the programmer may then rewrite his program as follows:

include "stdio.h"
print(‘hello world’);

And now the program will work. What did that magical line 'include stdio.h' do? Basically, it's the programmer saying the following:

“Ok, idiot. There’s this file called ‘stdio.h’. It will point you to another file called "stdio.c". stdio.h will list a bunch of functions' names (e.g., the function's name "f" in the above example and the fact that it receives one argument "x") and stdio.c will list their definitions (e.g., that f(x)=(x*1000+2 )/3). One of those functions is called ‘print’. Go read the functions in those two files, THEN read my code, and you’ll know what the function “print” is that I want to give ‘hello world’ to.”

“Print” is a real function in C++ (actually fprintf, but let's not split hairs: see this appearing in stdio.h). The above code will cause the string “hello world” to appear (be “printed”) on the standard output (your computer console). The function name “print(char* x)” is an example of a declaring code in an API (an Application Programming Interface) indicating the name of the function and what it receives as input (basically, a string of characters). Stdio.h and stdio.c, being a collection of such function declarations and their definitions, respectively, are an example of an API.

Incidentally, lest you think I'm making an unworthy analogy to JAVA with C++, note that the terms "declaration" and "header" are usually interchanged, e.g., as explained in the district court:

"The term "declaration" was used throughout trial to describe the headers (non-implementing code) for methods and classes. While "header" is the more technically accurate term, this order will remain consistent with the trial record and use "declaration" and "header" interchangeably."

You're reading this in a browser. Browsers have their own "compiler/interpreter" that reads its own language. The equivalent to the above code in that language is:

alert('hello world')

I tied that code to this button - click it:

Now slow down for a second. Think about what you just did.

With one, teeny, tiny line of code (whether "print" or "alert") amazing things happened. With "alert",

  1. your mouse talked to your input driver (Mouse:"I was clicked")
  2. your input driver talked to the operating system (Driver:"Hey, the mouse says he was clicked")
  3. the operating system talked to the browser (OS:"Hey browser, the mouse is at this position and says it was clicked")
  4. the browser talked to my code (Browser:"Is any code listening for a mouse click at this position?")
  5. my code talked to your browser (My code: "I am. Please send an alert")
  6. your browser talked to your operating system (Browser: "Can you put this alert on the screen for me?")
  7. the operating system talked to your display driver (OS: "Hey, the next time you refresh the display, draw this on there for me")
  8. and your display driver talked to your display (Driver: "Here's the next frame to draw")
That's the beauty and magic of APIs. I can wrap all manner of functionality into a simple, short set of calls. Attempting all that funcitonality from scratch every time we wrote a program would obviously be madness.

Now, it's important to appreciate that those function names "print" and "alert" interfacing into the API are both functional and fanciful.

That is, “printing” is what the function “print” does. Duh. That’s a functional description.

And yet, its author could instead have called it “lineout” or “send_to_console” or “to_screen”. Choosing to call it "print" was, to a certain extent, arbitrary.

Obviously, this is just one little line and one little choice, but in an API with thousands, maybe millions, of functions - that can be an awful lot of choices.

One word of caution: as mentioned there *is* some expressive character in the API per the above choice of "print." But that expressive character doesn't often manifest itself in the way, say, lines in a novel will.

20 lines of Jane Austen reflects 20 lines of distinct creative, expressive choices. 20 lines of API code may be a single creative choice varied 20 times (1: "#define constant A"; 2: "#define constant B"; "#define constant C"; etc.).

So when someone says "This header file has 1000 lines" it doesn't say much about the expressive content: it may be 1000 (or more) expressive choices, as in a novel . . . or it may be just one expressive choice repurposed 1000 times.

I've discussed OvG with lawyers and developers before and sometimes noticed that, even if they follow the fact pattern, court posture, etc. they tend to fall into one camp or the other, without really appreciating the objections of the other side.

I think that's because:

  • Until you've written a fair amount of software, it's difficult to appreciate what bits are expressive artistry and what bits are functional necessity (the attorneys' issue); and conversely
  • If you appreciate what bits are expressive artistry in software, it's hard to see how that relates to non-software precedent in the case law (the developers' issue)

For such folks, I've sometimes found it helpful to reference a close hypothetical analogue to the software fact pattern in OvG as follows:

In paper and pencil roleplaying games, like Dungeons and DragonsTM (DnD), you first create a character using a character creation sheet: Examples

Roleplaying games appeal to number crunchers. Accordingly, during creation, you may be given a finite number of "points", which you may allocate to different attributes of your character (e.g., with 50 points, 10 to intelligence, 20 to stamina, and 20 to charisma)

Another player, referred to as a Game Master, will then narrate an adventure for your character, presenting various challenges along the way. The outcome to your response to these challenges may then be determined using a combination of your character's attributes, the roll of one or more dice, and various game rules. E.g.,

Game Master: "You encounter a troll who swings at you with his fist."

You: "My dwarf attempts to dodge."

Game Master: "Very well. Calculate whether you succeeded using a roll of the dice, your dexterity attribute, and the dodge formula."

Some role playing games have very elaborate rules and character sheets. Let's consider the following hypothetical sheet, which combines a character sheet and rule book, as some games do.

The sheet has three sections:

  • The character attributes section, where you allocate the points for your character;
  • The calculator section, which provides instructions for calculating various intermediate values based upon your attributes and other inputs (these are like "implementing code" in an API); and
  • A "Character Fighting Move / Spell Index" indicating how to calculate the damage dealt when your character performs on of the maneuvers (this is like the "declaring code" in an API).

Note that the Index provides the move or spells name, the inputs it receives, and how those inputs are applied to various of the intermediate calculators to determine the resulting damage.

For example, if you choose to perform a "Flaming Sword" (name) maneuver, you must perform two dice rolls (inputs: roll_1, roll_2) and submit their results to the Sword Fighting (SF) and Spell Casting (SC) calculators, respectively, to determine how much damage you did.

Thus, the index (API declarations) serve as a reference to the more substantive computational operations of the calculators (API implementing code)

Thus, under our analogy, the "Character Fighting Move / Spell Index" corresponds to the JAVA API declaring code (what Google copied), the calculators correspond to the implementing code (what Google rewrote), and the sheet as a whole corresponds to the JAVA language as a whole.

It should be clear that, despite its having some functional character, as a whole such a sheet is copyrightable. Even in this barebones example, the selection and arrangement of elements are expressive. Similarly, no one denies that JAVA, as a whole, is copyrightable.

I've found that this helps convey some of the existential angst at play in the case. Basically, after Sun/Oracle's character sheet enjoyed wild popularity in the fantasy game domain, Google decided to create its own sheet for a sci-fi domain, rewriting everything in the sheet, except for the Index, which they copied wholesale, because they wanted players to avail themselves of the familiarity they'd gained using Oracle's sheet.

This creates a policy kerfuffle. Is this like someone creating a derivative Harry Potter novel, reusing the characters? I mean, the Index has some fanciful name selections. They're not entirely fanciful, though - there's only so many ways you can refer to a "love spell." So is it more like someone reusing a functional, standardized protocol? If we say this is fair use, how will that affect the creation of new character sheets? If we say that it isn't fair use, how will that affect the creation of new character sheets?

Again, this is just a conceptual analogue, but the issues are close enough to OvG to convey the complexities involved.

Judge Boudin's concurrence in the 1995 Lotus v. Borland is essentially the prequel to OvG. There's an entry for Lotus v. Borland in the case notes. Basically, the First Circuit found a "command hierarchy" menu interface to be uncopyrightable (it was pretty barebones apparently, e.g., "print" printed, etc.).

I'm lacking time to walk through it with you at present and so will copy it below for reference for now.

I highlight what I refer to as Boudin's Scylla and Charybdis below.

Concurrence follows. BOUDIN, Circuit Judge, concurring. The importance of this case, and a slightly different emphasis in my view of the underlying problem, prompt me to add a few words to the majority's tightly focused discussion.


Most of the law of copyright and the "tools" of analysis have developed in the context of literary works such as novels, plays, and films. In this milieu, the principal problem--simply stated, if difficult to resolve--is to stimulate creative expression without unduly limiting access by others to the broader themes and concepts deployed by the author. The middle of the spectrum presents close cases; but a "mistake" in providing too much protection involves a small cost: subsequent authors treating the same themes must take a few more steps away from the original expression.

The problem presented by computer programs is fundamentally different in one respect. The computer program is a means for causing something to happen; it has a mechanical utility, an instrumental role, in accomplishing the world's work. Granting protection, in other words, can have some of the consequences of patent protection in limiting other people's ability to perform a task in the most efficient manner. Utility does not bar copyright (dictionaries may be copyrighted), but it alters the calculus.

Of course, the argument for protection is undiminished, perhaps even enhanced, by utility: if we want more of an intellectual product, a temporary monopoly for the creator provides incentives for others to create other, different items in this class. But the "cost" side of the equation may be different where one places a very high value on public access to a useful innovation that may be the most efficient means of performing a given task. Thus, the argument for extending protection may be the same; but the stakes on the other side are much higher.

It is no accident that patent protection has preconditions that copyright protection does not--notably, the requirements of novelty and non-obviousness--and that patents are granted for a shorter period than copyrights. This problem of utility has sometimes manifested itself in copyright cases, such as Baker v. Selden, 101 U.S. 99, 25 L. Ed. 841 (1879), and been dealt with through various formulations that limit copyright or create limited rights to copy. But the case law and doctrine addressed to utility in copyright have been brief detours in the general march of copyright law.

Requests for the protection of computer menus present the concern with fencing off access to the commons in an acute form. A new menu may be a creative work, but over time its importance may come to reside more in the investment that has been made by users in learning the menu and in building their own mini-programs--macros--in reliance upon the menu. Better typewriter keyboard layouts may exist, but the familiar QWERTY keyboard dominates the market because that is what everyone has learned to use. See P. David, CLIO and the Economics of QWERTY, 75 Am. Econ. Rev. 332 (1985). The QWERTY keyboard is nothing other than a menu of letters.

Thus, to assume that computer programs are just one more new means of expression, like a filmed play, may be quite wrong. The "form"--the written source code or the menu structure depicted on the screen--look hauntingly like the familiar stuff of copyright; but the "substance" probably has more to do with problems presented in patent law or, as already noted, in those rare cases where copyright law has confronted industrially useful expressions. Applying copyright law to computer programs is like assembling a jigsaw puzzle whose pieces do not quite fit.

All of this would make no difference if Congress had squarely confronted the issue, and given explicit directions as to what should be done. The Copyright Act of 1976 took a different course. While Congress said that computer programs might be subject to copyright protection, it said this in very general terms; and, especially in § 102(b), Congress adopted a string of exclusions that if taken literally might easily seem to exclude most computer programs from protection. The only detailed prescriptions for computers involve narrow issues (like back-up copies) of no relevance here.

Of course, one could still read the statute as a congressional command that the familiar doctrines of copyright law be taken and applied to computer programs, in cookie cutter fashion, as if the programs were novels or play scripts. Some of the cases involving computer programs embody this approach. It seems to me mistaken on two different grounds: the tradition of copyright law, and the likely intent of Congress.

The broad-brush conception of copyright protection, the time limits, and the formalities have long been prescribed by statute. But the heart of copyright doctrine--what may be protected and with what limitations and exceptions--has been developed by the courts through experience with individual cases. B. Kaplan, An Unhurried View of Copyright 40 (1967). Occasionally Congress addresses a problem in detail. For the most part the interstitial development of copyright through the courts is our tradition.

Nothing in the language or legislative history of the 1976 Act, or at least nothing brought to our attention, suggests that Congress meant the courts to abandon this case-by-case approach. Indeed, by setting up § 102(b) as a counterpoint theme, Congress has arguably recognized the tension and left it for the courts to resolve through the development of case law. And case law development is adaptive: it allows new problems to be solved with help of earlier doctrine, but it does not preclude new doctrines to meet new situations.


In this case, the raw facts are mostly, if not entirely, undisputed. Although the inferences to be drawn may be more debatable, it is very hard to see that Borland has shown any interest in the Lotus menu except as a fall-back option for those users already committed to it by prior experience or in order to run their own macros using 1-2-3 commands. At least for the amateur, accessing the Lotus menu in the Borland Quattro or Quattro Pro program takes some effort.

Put differently, it is unlikely that users who value the Lotus menu for its own sake--independent of any investment they have made themselves in learning Lotus' commands or creating macros dependent upon them--would choose the Borland program in order to secure access to the Lotus menu. Borland's success is due primarily to other features. Its rationale for deploying the Lotus menu bears the ring of truth.

Now, any use of the Lotus menu by Borland is a commercial use and deprives Lotus of a portion of its "reward," in the sense that an infringement claim if allowed would increase Lotus' profits. But this is circular reasoning: broadly speaking, every limitation on copyright or privileged use diminishes the reward of the original creator. Yet not every writing is copyrightable or every use an infringement. The provision of reward is one concern of copyright law, but it is not the only one. If it were, copyrights would be perpetual and there would be no exceptions.

The present case is an unattractive one for copyright protection of the menu. The menu commands (e.g., "print," "quit") are largely for standard procedures that Lotus did not invent and are common words that Lotus cannot monopolize. What is left is the particular combination and sub-grouping of commands in a pattern devised by Lotus. This arrangement may have a more appealing logic and ease of use than some other configurations; but there is a certain arbitrariness to many of the choices.

If Lotus is granted a monopoly on this pattern, users who have learned the command structure of Lotus 1-2-3 or devised their own macros are locked into Lotus, just as a typist who has learned the QWERTY keyboard would be the captive of anyone who had a monopoly on the production of such a keyboard. Apparently, for a period Lotus 1-2-3 has had such sway in the market that it has represented the de facto standard for electronic spreadsheet commands. So long as Lotus is the superior spreadsheet--either in quality or in price--there may be nothing wrong with this advantage.

But if a better spreadsheet comes along, it is hard to see why customers who have learned the Lotus menu and devised macros for it should remain captives of Lotus because of an investment in learning made by the users and not by Lotus. Lotus has already reaped a substantial reward for being first; assuming that the Borland program is now better, good reasons exist for freeing it to attract old Lotus customers: to enable the old customers to take advantage of a new advance, and to reward Borland in turn for making a better product. If Borland has not made a better product, then customers will remain with Lotus anyway.

BEGIN: Boudin's Scylla and Charybdis

Thus, for me the question is not whether Borland should prevail but on what basis. Various avenues might be traveled, but the main choices are between holding that the menu is not protectable by copyright [THE SCYLLA] and devising a new doctrine that Borland's use is privileged [THE CHARYBDIS]. No solution is perfect and no intermediate appellate court can make the final choice.

[THE SCYLLA]To call the menu a "method of operation" is, in the common use of those words, a defensible position. After all, the purpose of the menu is not to be admired as a work of literary or pictorial art. It is to transmit directions from the user to the computer, i.e., to operate the computer. The menu is also a "method" in the dictionary sense because it is a "planned way of doing something," an "order or system," and (aptly here) an "orderly or systematic arrangement, sequence or the like." Random House Webster's College Dictionary 853 (1991).

[THE CHARYBDIS] A different approach would be to say that Borland's use is privileged because, in the context already described, it is not seeking to appropriate the advances made by Lotus' menu; rather, having provided an arguably more attractive menu of its own, Borland is merely trying to give former Lotus users an option to exploit their own prior investment in learning or in macros. The difference is that such a privileged use approach would not automatically protect Borland if it had simply copied the Lotus menu (using different codes), contributed nothing of its own, and resold Lotus under the Borland label.

The closest analogue in conventional copyright is the fair use doctrine. E.g., Harper & Row, Publishers, Inc. v. Nation Enters., 471 U.S. 539, 85 L. Ed. 2d 588, 105 S. Ct. 2218 (1985). Although invoked by Borland, it has largely been brushed aside in this case because the Supreme Court has said that it is "presumptively" unavailable where the use is a "commercial" one. See id. at 562. But see Campbell v. Acuff-Rose Music, Inc., 127 L. Ed. 2d 500, 114 S. Ct. 1164, 1174 (1994). In my view, this is something less than a definitive answer; "presumptively" does not mean "always" and, in any event, the doctrine of fair use was created by the courts and can be adapted to new purposes.

But a privileged use doctrine would certainly involve problems of its own. It might more closely tailor the limits on copyright protection to the reasons for limiting that protection; but it would entail a host of administrative problems that would cause cost and delay, and would also reduce the ability of the industry to predict outcomes. Indeed, to the extent that Lotus' menu is an important standard in the industry, it might be argued that any use ought to be deemed privileged.

END: Boudin's Scylla and Charybdis

In sum, the majority's result persuades me and its formulation is as good, if not better, than any other that occurs to me now as within the reach of courts. Some solutions (e.g., a very short copyright period for menus) are not options at all for courts but might be for Congress. In all events, the choices are important ones of policy, not linguistics, and they should be made with the underlying considerations in view.

Should we choose the Scylla? The Charybdis? Or sail around entirely and let Congress decide?

This is the question OvG will address.

One of the biggest challenges in reviewing a fair use analysis is to have the constellation of disparate decisions of varying precedential effect at your command.

The Copyright Office provides a fair use case index. However, it doesn't break apart the PNAE analyses, doesn't cite actual opinion language (i.e., the raw logical propositions), or include some of the copyrightability opinions relevant here.

Accordingly, to ensure that my analysis included all the precedential propositions, I wrote my own interactive, sortable index.

I decided to provide it here for your consideration, but recognize that it was highly informal, being originally for my own personal use.

There's roughly ~50 cases outside the OvG opinions in this constellation..

Clicking up and down arrows will change the sort order. Clicking the checkbox will show/hide the column.

Fair Use Analysis

Sun Microsystems released the first public implementation of JAVA in 1996 (though James Gosling, Mike Sheridan, and Patrick Naughton had been working on it since 1991).

In August 2005, Google acquired Android, Inc., as part of a plan to develop a smartphone platform. Developer adoption was critical to capturing market share from competing phones, such as Apple's iPhone, that would also be released in the near future (the iPhone came out in 2007). Since developers were already familiar with JAVA, Google decided to use the Java language for the Android platform.

Accordingly, in late 2005, Google approached Sun about the possibility of taking a license to use and adapt the entirety of JAVA for mobile devices. They also discussed a possible co-development partnership deal with Sun under which Java technology would become an opensource part of the Android platform, adapted for mobile devices. Despite negotiating for over several months, they were unable to reach a deal.

In 2007, Google Inc., announced its Android software platform for mobile devices.

The first phones go on sale in 2008.

In 2008, the Java API had grown from its original eight packages to 166 packages with over six hundred classes with over six thousand methods. Google's Android platform replicated the same package, method and class names, definitions and parameters of the 37 Java API packages from the Java 2SE 5.0 platform. The district court provided a nice example:

package java.lang;
public class Math {
public static int max (int x, int y) {
if (x > y) return x ;
else return y ;
// Declares package java.lang [COPIED]
// Declares class Math [COPIED]
// Declares method max [COPIED]
// Implementation, returns x or [REWRITTEN]
// Implementation, returns y [REWRITTEN]
// Closes method
// Closes class

Though typically there are many more than the two lines of implementing code. Indeed, Google wrote six thousand-plus subroutines (methods) and six hundred- plus classes of implementing code. The Android platform has its own API, these 37 packages being part of the 168 total packages in that API.

While Google's packages retained the groupings from the original, the actual sequences of methods in the listings are different. That is, the sequence of methods in the class Math in Android is different from the sequence in the same class in Java, although all of the methods in the Java version can be found somewhere in the Android version, at least as shown in their respective listings.

In 2009, Dr. Bloch (who had worked at Sun from August 1996 through July 2004, eventually holding the title of "distinguished engineer") worked on Google's Android project for approximately one year. While working on the Android team, Dr. Bloch contributions to the Android platform included a nine-line rangeCheck function from JAVA, which appeared in a class containing 3,179 lines of code. When discovered, the rangeCheck lines were removed.

Oracle Corporation acquired Sun Microsystems on January 27, 2010.

Thus, in summary, the three items at issue were, with the first two being, effectively, discarded:

  1. 8 security files, which were copied, but did not appear in Android being instead used merely for internal testing (thus, the OvG analysis doesn't really focus upon these)
  2. The nine-line rangeCheck code, which was removed (again, being de minimis, the OvG analysis doesn't really focus upon this)
  3. Declaratory code as described above for 37 Packages (in contrast, the controversy focuses here)
    • Specifically:
      1. java.awt.font
      2. java.beans
      3. java.io
      4. java.lang
      5. java.lang.annotation
      6. java.lang.ref
      7. java.lang.reflect
      8. java. net
      9. java.nio
      10. java.nio.channels
      11. java.nio.channels.spi
      12. java.nio.charset
      13. java.nio.charset.spi
      14. java.security
      15. java.security.acl
      16. java.security.cert
      17. java.security.interfaces
      18. java.security.spec
      19. java.sql
      20. java.text
      21. java.util
      22. java. util.jar
      23. java.util.logging
      24. java.util.prefs
      25. java.util.regex
      26. java.util.zip
      27. javax.crypto
      28. javax.crypto.interfaces
      29. javax.crypto.spec
      30. javax.net
      31. javax.net.ssl
      32. javax.security.auth
      33. javax. security.auth.callback
      34. javax.security.auth.login
      35. javax.security.auth.x500
      36. javax.security.cert
      37. javax.sql
    • I couldn't get access to the CD containing the exact code submitted into evidence, but I believe you can review the same (or at least very similar) declarations here. For example, here are the declaration items for "java.lang.annotations". Similarly, you can reference the JDK for lines, as in this example. Obviously, ignore the comments in the JDK - it's just the declaring code that was copied.

In 2010, Oracle files suit against Google in the Northern California District alleging that Google’s Android mobile operating system infringed Oracle’s patents and copyrights. Incidentally, CAFC hears the appeal and not the 9th circuit because of the patent claim - even after those claims are dropped. This is another benefit to having a large patent portfolio, as it can help you select whether you want CAFC to hear the appeal or not.

It is, generally, true that the Supreme Court's (SCOTUS) opinion supercedes the underlying opinions. But where SCOTUS is intentionally or unintentionally vague in its reasoning, we should consider lower court decisions for reference. Posture / context affects interpretation.

As of 04/2021, there are generally six opinions that are worth our consideration, broken into two threads:

  1. The "Copyrightablity" thread:
    1. The 2012 California Distrcit Court Decision on Copyrightability
    2. The 2014 Court of Appeals for the Federal Circuit Decision on Copyrightability
    3. The 2015 Supreme Court Decision to Decline Consideration of the Copyrightability Question
  2. The "Fair Use" thread:
    1. The 2016 California Distrcit Court Decision on Fair Use
    2. The 2018 Court of Appeals for the Federal Circuit Decision on Fair Use
    3. The 2021 Supreme Court Decision on Fair Use

1. 2012 District Court Decision on Copyrightability
Supreme Court of the United States
Court of Appeals for the Federal Circuit
California District Court
Copyrightable? NO
(We're here)

Summary: Relying heavily on (its own interpretation of) Lotus v. Borland, the Court concludes that the API declaring code is not copyrightable.

Oracle Appeals
2. 2014 CAFC Decision on Copyrightability
Supreme Court of the United States
Court of Appeals for the Federal Circuit
Copyrightable? YES
(We're here)
California District Court

Summary: CAFC overrules the district court.

Google Appeals
3. 2015 SCOTUS Declines to hear appeal on Copyrightability
Supreme Court of the United States
Court of Appeals for the Federal Circuit
California District Court

Summary: SCOTUS declines to hear the case.

Sent back to District Court
4. 2016 District Court Decision(s) on Fair Use
Supreme Court of the United States
Court of Appeals for the Federal Circuit
California District Court
Fair Use? YES
(We're here)

Summary: The district court jury hears the entire fair use question and finds fair use.

Oracle Appeals
5. 2018 CAFC Decision on Fair Use
Supreme Court of the United States
Court of Appeals for the Federal Circuit
Fair Use? NO
(We're here)
California District Court

Summary: CAFC overrules as a matter of law, determining that juries only determine historical facts.

Google Appeals
(With cross-appeal raising copyrightability again)
6. 2021 Supreme Court Decision on Fair Use
Supreme Court of the United States
Fair Use? YES
(We're here)
Court of Appeals for the Federal Circuit
California District Court

Summary: SCOTUS overrules, finding fair use as a matter of law and noting that the 7th amendment doesn't guarantee a jury's consideration of fair use. SCOTUS doesn't address the copyrightability question.

Full Opinion

The Court first provides a very methodical walkthrough of API operation.quote If you've read the API overview in this writeup, it basically covers the same terriroty, but focusing on JAVA.

After that, Judge Alsup gives a very thorough review of the case lawquote, specifically touching upon:

  • Baker v. Selden, 101 U.S. 99, 25 L. Ed. 841, 1880 Dec. Comm'r Pat. 422 (1879)
  • The Computer Age and Section 102(b) of the 1976 Act.
  • Apple Computer, Inc. v. Formula Intern. Inc., 725 F.2d 521, 522-25 (9th Cir. 1984).
  • Apple Computer, Inc. v. Microsoft Corp., 35 F.3d 1435, 1445 (9th Cir. 1994).
  • Computer Associates International, Inc. v. Altai, Inc., 982 F.2d 693 (2d Cir. 1992)
  • Sega Enters., Ltd. v. Accolade, Inc., 977 F.2d 1510, 1524-25 (9th Cir. 1992).
  • Gates Rubber Co. v. Bando Chemical Industries, Ltd., 9 F.3d 823 (10th Cir. 1993).
  • Whelan Associates, Inc. v. Jaslow Dental Laboratory, Inc., 797 F.2d 1222 (3d Cir. 1986).
  • Lotus Development Corp. v. Borland International, Inc., 49 F.3d 807 (1st Cir. 1995).
  • Hutchins v. Zoll Medical Corp., 492 F.3d 1377 (Fed. Cir. 2007)
  • Johnson Controls, Inc. v. Phoenix Control Sys., Inc., 886 F.2d 1173, 1175 (9th Cir. 1989).
  • Feist Publications, Inc. v. Rural Telephone Services Co., Inc., 499 U.S. 340, 111 S. Ct. 1282, 113 L. Ed. 2d 358 (1991),
  • Brown Bag Software v. Symantec Corp., 960 F.2d 1465 (9th Cir. 1992),
  • Atari Games Corp. v. Nintendo of America Inc., 975 F.2d 832 (Fed. Cir. 1992),
  • Sony Computer Entertainment, Inc., v. Connectix Corporation, 203 F.3d 596 (9th Cir. 2000)
Again, you can find links to the full opinions in the case summaries.

Following this comprehensive survey, the Court then announces that it will use the following precedential doctrines in its analysis quote :

  • The merger doctrine (for the class and method names)
  • The uncopyrightability of short phrases (for the class and method names)quote
  • Section 102(b) and the uncopyrightability of "any idea, procedure, process, system, method of operation or concept" (this is apparently applied as part of the "command structure" disucssion - see the below quote for Feist)
  • Feist as "we should not yield to the temptation to find copyrightability merely to reward an investment made in a body of intellectual property." (actually, after this announcement, there's no explicit reference to Feist, but it's likely implied in the discussion of 102 herequote)

The Court applies those doctrines in an ordered fashion:

  • First, the Court addresses "methods" and "classes"quote in the API:
    • At the method level, the Court finds the merger and short phrase doctrines prevents copyrightability of the methods. quote
      • (For clarity, if you're a developer and confused by the Court's apparent conflation between method names and method implementations, that's not you, that's the Court and something CAFC will address in the next opinion)
    • At the class level, the Court similarly appears to assert that class names are uncopyrightable under the merger and short phrase doctrines quote
  • Second, the Court addresses the orgnization of the API as a whole quote
    • Previously in its analysis, when discussing Baker v. Selden, the Court suggested this was better suited for patent coverage quote
    • The Court distinguishes the taxonomy which was found copyrightably in America Dental on the principle that this is a "command structure." quote
    • Interoperability: the Court relies heavily upon the APIs function for interoperability as a basis for construing it as a "command structure"quote with clear echoes of Lotus v. Borland.
  • Ultimately, the Court concludes that, even if the declaring code is a "taxonomy", as a "command structure", it is not copyrightable.
    This explains why American Dental Association v. Delta Dental Plans Association, 126 F.3d 977 (7th Cir. 1997), is not controlling. Assuming arguendo that a taxonomy is protectable by copyright in our circuit, see Practice Mgmt. Info. Corp. v. Am. Med. Ass'n, 121 F.3d 516 (9th Cir. 1997), the taxonomy in ADA had nothing to do with computer programs. It was not a system of commands, much less a system of commands for a computer language.
  • After concluding, the court walks back a bit, tailoring its reasoning to the particular posture of this case.quote

  • This opinion starts strong, but I'm afraid its reasoning grows progressively more vague (CAFC will be similarly underwhelmed).

    For example, the opinion says:

    In this regard, the Ninth Circuit decisions in Sega and Sony, although not on all fours, are close analogies. Under these two decisions, interface procedures required for interoperability were deemed "functional requirements for compatibility" and were not copyrightable under Section 102(b). Both decisions held that interface procedures that were necessary to duplicate in order to achieve interoperability were functional aspects not copyrightable under Section 102(b).

    Really? Is that what those cases held? The reasoning here is vague. The phrase "functional requirements for compatibility" doesn't appear in Sony. It does appear in Sega. The court quoted Sega above in the caselaw section, saying:

    In summary, the record clearly establishes that disassembly of the object code in Sega's video game cartridges was necessary in order to understand the functional requirements for Genesis compatibility. The interface procedures for the Genesis console are distributed for public use only in object code form, and are not visible to the user during operation of the video game program. Because object code cannot be read by humans, it must be disassembled, either by hand or by machine. Disassembly of object code necessarily entails copying. Those facts dictate our analysis of the second statutory fair use factor. If disassembly of copyrighted object code is per se an unfair use, the owner of the copyright gains a de facto monopoly over the functional aspects of his work — aspects that were expressly denied copyright protection by Congress..
    Sega v. Accolade
    But from this quote alone, it's not clear what conclusion the District Court expects us to draw. Specifically, how does a holding that "disassembly of specific, functional, assembled code is a fair use" imply that "this specific uncompiled declaratory code cannot be copyrighted"? The opinion would seem to respond "because it's a 'command structure'", but assuming that phrase is being drawn from Lotus v. Borland, Lotus held that the "the Lotus menu command hierarchy is an uncopyrightable 'method of operation'" not that "command structures" are inherently uncopyrightable.

    These discrepancies are things CAFC will pick on in their opinion.

  • I think it's very important in all these decisions, to keep in mind how narrowly tailored they are to the facts at hand.

    Here, the court acknowledged several times that less than all the code at issue had been considered:

    At trial, however, neither side presented evidence from which we can now say that the same is true for all the other hundreds of classes at issue. Therefore, it is impossible to say on this record that all of the classes and their contents are typical of such classes and, on this record, this order rejects Google's global argument based on scenes a faire . . . The Court has not compared all six-hundred-plus classes. Nor has any witness or counsel so far on the record. Oracle does not, however, contend that the actual sequences would track method-for-method and it has not so proven. This detailed observation, however, does not change the fact that all of the methods in the Java version can be found somewhere in the Android version, classified under the same classes.

    There's no real discussion digging into the API structure. There's no real application of the Abtraction Filtration Comparison test (something CAFC will also take issue with). APIs often involve variations on a creative selection. What were those selections? How were they varied?

  • Judge Alsup seems to put a lot of faith in patents as being a suitable alternative here.
    Yes, it is creative. Yes, it is original. Yes, it resembles a taxonomy. But it is nevertheless a command structure, a system or method of operation — a long hierarchy of over six thousand commands to carry out pre-assigned functions. For that reason, it cannot receive copyright protection — patent protection perhaps — but not copyright protection.
    However, after the Alice v. CLS Bank SCOTUS case, it's very difficult to acquire patent coverage on generalized structures like these.
  • The court keeps using the phrase "command structure" without saying where it comes from - we're left to infer that it's from Lotus v. Borland. quote

Full Opinion

  • Judges O'Malley, Plager, and Taranto at CAFC reiterate the factual history, the precedings below, and then reference various authority they'll use for their analysis including:
    • Atari Games Corp. v. Nintendo of Am., Inc., 975 F.2d 832, 838
    • Feist Publ’ns, Inc. v. Rural Tel. Serv. Co., 499 U.S. 340, 358 (1991).
    • Golan v. Holder, 132 S. Ct. 873, 890 (2012)
    • Baker v. Selden, 101 U.S. 99, 101 (1879)
    • Computer Assocs. Int’l v. Altai, 982 F.2d 693, 704-05 (2d Cir. 1992)
    • Johnson Controls, Inc. v. Phoenix Control Sys., Inc., 886 F.2d 1173,
    • Lexmark Int’l, Inc. v. Static Control Components, Inc., 387 F.3d 522, 533 (6th Cir. 2004)
    • Lotus Dev. Corp. v. Borland Int’l, 49 F.3d 807
    • Whelan Assocs., Inc. v. Jaslow Dental Lab., Inc.
  • CAFC provides a slightly more thorough discusion of literal and non-literal copying than the district court.quote
  • CAFC then discusses the Abstraction-Filtration-Comparison (AFC) test, noting that the circuits are in accord on the first step, but not the second.
  • The court notes in a footnote that, for literal copying, one goes straight to the filtration analysis quote
  • CAFC then notes that, while the district court referenced the AFC test, the district court didn't actually apply it. quote
  • CAFC then begins its analysis in earnest:
    • Declaring Source Code
      • Oracle Waiver: Initially, the court clarifies that Oracle didn't waive its arguments as to the delcaring codequote. CAFC then considers the district court's analysis of the declaring code under the Merger and the Short Phrases doctrines.
      • Merger
        • First, CAFC consider merger irrelevant quote as "“unique arrangement of computer program expression . . . does not merge with the process so long as alternate expressions are available.” Atari, 975 F.2d at 840." Below I mentioned the expression discussion was confusing - CAFC picks that apart herein their opinion.quote. CAFC notes that, if there hadn't been alternative methods for expressing the method and class names, then maybe merger would apply quote. So CAFC reverses Judge Alsup's merger analyis.
      • Short Phrases
        • CAFC next says Judge Alsup's short phrases analysis was incomplete, as the issue is "not whether the work at issue contains short phrases—as literary works often do" but rather "whether those phrases are creative."
        • Further, the district court should have looked at combinations of phrases instead of dissecting into individual words.quote
      • Scenes a Faire
        • CAFC reiterates that the district court rejected Google’s reliance on the scenes a faire doctrine. It did so in a footnote, finding that Google had failed to present evidence to support the claim that either the grouping of methods within the classes or the code chosen for them “would be so expected and customary as to be permissible under the scenes a faire doctrine.”
        • CAFC says Google cannot rely on Scenes a Faire:
          1. As it's part of the inringement analysis, not not the copyrightability analysis
          2. Google has not objected to the district court’s conclusion that Google failed to make a sufficient factual record to support its contention that the groupings and code chosen for the 37 Java API packages were driven by external factors or premised on features that were either commonplace or essential to the idea being expressed
          3. Google is trying to apply scenes a faire to the copier instead of upon the creator ("the focus of the scenes a faire doctrine is on the circumstances presented to the creator, not the copier.") quote
    • The Structure Sequence Operation of the API
      • As I noted above, CAFC likewise infers that Judge Alsup had Lotus in mind when referencing "command structures" quote
      • CAFC then distinguihes the facts of Lotus from the facts at issue here, but, more importanly, explains that Lotus shouldn't be used for the propositions the district court used, as doing so contradicts Ninth Circuit lawquote
      • Importantly, CAFC notes that just because code is functional doesn't mean its expressive character is protectable quote
      • Again, the emphasis is on the variation available to the designers quote. Accordingly, CAFC finds the SSO to be copyrightable.
    • Interoperability is Irrelevant to Copyrightability
      • CAFC distinguishes the tangential refenreces to interoperability in the Sega and Sony decisions quote and notes that, to the extent those decisions apply here, it is with regard to fair use
      • Importantly, CAFC reemphasizes that the infringer will often be subject to constraints, be they in hardware, in the ecosystem, etc. But constraints upon the infringer are not the constraints relevant to copyrightability analysis - it's the constraints upon the author. quote
      • Accordingly, CAFC determines that interoperability is irrelevant to the copyrightability analysis. quote
      • Also, importantly, CAFC reiterates that popularity of a standard is not a question for copyrightability quote
  • CAFC then states that a Fair Use analysis is premature on the available record, and that the cases should go back down for further factual assessment.quote
    • Indeed, CAFC acknowledges that some of the arguments mistakenly applied to copyrightability may bear upon fair use.quote
  • Google Cross Appeal
    • Google cross appealed regarding rangeCheck and the 8 decompiled files, but CAFC affirms Oracle's JMOL motions quote
  • Google Policy Arguments
    • Google reiterates the district court's preference for patent coverage. "Google argues that “[a]fter Sega, developers could no longer hope to protect [software] interfaces by copyright . . . Sega signaled that the only reliable means for protecting the functional requirements for achieving interoperability was by patenting them." CAFC dismisses the argument. quote

  • Scenes a Faire probably WILL come up in many API cases, so it was rather odd that it got set aside here.
  • Like the district court, Google and Amici argue that patent law is the solution:
    Many of Google’s arguments, and those of some amici, appear premised on the belief that copyright is not the correct legal ground upon which to protect intellectual property rights to software programs; they opine that patent protection for such programs, with its insistence on non-obviousness, and shorter terms of protection, might be more applicable, and sufficient.

    In a dark humor sort of way, this is rather funny. CAFC hears Alice v. CLS Bank appeals all the time invalidating patent coverage for general software concepts, so for a party to say "just give them patent law, that should do it" is rather amusing.

  • This was a good, thorough decision.
  • SCOTUS denies Google's appeal from the CAFC opinion.
  • There are actually some interesting points in the Solicitor General's brief and some other filings, but in the interest of economy / time, I'll hold off on writing those up here until folks express an interest. To a first order, you can understand OvG without reviewing them in detail.

Full Opinion

The case returns to the district court and proceeds to trial on fair use. *All aspects* of the fair use analysis are given to the jury. After roughly one week of evidence and several days of deliberations, the jury finds that Google's use of the declaring lines of code and the SSO of the 37 API packages was FAIR USE.

The district court's opinion then considers the reasonableness of the jury's conclusions.


This factor is broken into commerciality and transformation sub-factors.

Commercialityquote - The court infers that the jury either:

  • Considered the use purely commercial
  • OR
  • Considered the use to include non-commercial uses.


open-source mobile operating system for smartphones was transformative Indeed, a “leading treatise on this topic has lamented the frequent misuse of the transformation test, complaining that it has become a conclusory label which is ‘all things to all people.’” Id. (quoting Melville B. Nimmer & David Nimmer, 4 Nimmer on Copyright § 13.05[A][1][b], 13168-70 (2011)).

Bad Faith:quote quote

The Court basically concludes that the jury had adequate information / instructions, including custom, to consider the issue in Google's favor.

The district court concludes that the jury eitherquote
Could reasonably have concluded that the declaring code was highly creative (but gave this factor less weight)
Could reasonably have gone the other way and concluded that the declaring code was not highly creative.
The district court infers that the Jury concludedquote
Could reasonably have found that Google duplicated the bare minimum of the 37 API packages, just enough to preserve inter-system consistency in usage
The district court infers that the Juryquote . . .
. . . could reasonably have found that use of the declaring lines of code (including their SSO) in Android caused no harm to the market for the copyrighted works, which were for desktop and laptop computers.

Summary (N.B., this isn't treated like a "scorecard", the factors may not be given equal weight; still, courts usually phrase their reasoning so a majority of the factors come out on one side):

urpose of the Use: Either Favors Fair Use or Neutral
ature of the copyrighted work: Either Favors Fair Use or Doesn't Favor Fair Use
mount of the Work Used: Favors Fair Use
ffect Upon the Market for the Work: Favors Fair Use

  • As you may have inferred, approaching fair use determinations in this manner may seem a bit problematic. The jury is basically operating as a "black box": we gave them some questions, they gave us some answers. Which factors did the jury find most relevant? Dunno. Why? Dunno. Which facts were considered in relation to those elements? Dunno. Why? Dunno.

    Asking the district court judge to "guesstimate" the jury's reasoning is unfair and doesn't really help the precedential record much.

    Typically, questions of "fact" go to the jury and questions of "law" are decided by the judge. I'm not sure why the parties agreed to send everything to the jury here (I'll put an update here if I find an explanation in the docket). You can send "special interrogatories" to the jury, asking them to explain their reasoning (I did ask a litigator friend and he said that sometimes the parties won't submit interrogatories if they're both confident that the jury is going to come out on their side and they don't want to belabor the proceedings). Absent such interrogatories, the law presumes the existence of findings necessary to support the verdict the jury reached.

    CAFC provides a bit of clarification in its opinion. quote

    Fortunately, CAFC and the Supreme Court will establish that juries only determine historical facts and the judge will determine the inferences to be drawn from those facts. So the precedential record should be better in future cases.

Full Opinion

Oracle appeals the district court's determination of fair use to the CAFC.

Wrapped within the fair use question are three preliminary questions concerning juries and fair use (this is actually part of an ongoing application of the Supreme Court's U.S. Bank decision, trying to figure out the jury's role in various mixed questions of law and factquote), answered as follows:quote:

  1. Query #1: Which aspects of fair use are legal and which are factual?
    • Answer #1:
      • The jury's role is limited to determining disputed "historical facts," not the inferences or conclusions to be drawn from those facts (note this is similar to obviousness in patent cases)quote
  2. Query #2: Under what standard are the factual questions reviewed?
    • Answer #2:quote
      • "Court's application of the correct legal standard to the fair use inquiry": reviewed de novo (basically meaning, they redo the analysis)
      • "whether the findings relating to any relevant historical facts were correct are questions": reviewed with deference (not as fresh as de novo)
      • "whether the use at issue is ultimately a fair one": reviewed de novo.
  3. Query #3: While the jury may determine subsidiary factual questions, are there any aspects of fair use which the jury make the final decision?
    • Answer #3:
      • No. Any jury findings other than those concerning historical fact are merely "advisory."quote

With that out of the way, CAFC applies these conclusions and standard of review to the fair use analysis of the district court. As discussed above, the district court sent everything to the jury.quote CAFC finds the district court's analysis deficient as follows:

CAFC's analysis parallels the district court's, but with applications of precedent rather than inferences from the jury.quote
"It is undisputed that the API packages 'serve the same function in both works . . . The fact that Google created exact copies of the declaring code and SSO and used those copies for the same purpose as the original material “seriously weakens [the] claimed fair use.”
Commercial Usequote
CAFC discounts the district court's reliance upon Android being open source in its commercial use analysis.
"Giving customers 'for free something they would ordinarily have to buy' can constitute commercial use."
Transformative Usequote
Directly contravening the district court's conclusion, the court finds the use is not transformative. CAFC identifies the precedent in Sony as germane to this factorquote
"Google’s use of the API packages is not transformative as a matter of law because: (1) it does not fit within the uses listed in the preamble to § 107; (2) the purpose of the API packages in Android is the same as the purpose of the packages in the Java platform; (3) Google made no alteration to the expressive content or message of the copyrighted material; and (4) smartphones were not a new context.
. . .
That Google wrote its own implementing code is irrelevant to the question of whether use of the APIs was transformative. As we noted in the prior appeal, “no plagiarist can excuse the wrong by showing how much of his work he did not pirate."

Recall, that the district court considered that hte code was used in smartphones, rather than in desktops or laptops, to be transformative. However, here, CAFC says:

"Google’s primary argument on appeal is that Android is transformative because Google incorporated the declarations and SSO of the 37 API packages into a new context— smartphones. But the record showed that Java SE APIs were in smartphones before Android entered the market. Specifically, Oracle presented evidence that Java SE was in SavaJe mobile phones and that Oracle licensed Java SE to other smartphone manufacturers, including Danger and Nokia. Because the Java SE was already being used in smartphones, Google did not “transform” the copyrighted material into a new context and no reasonable jury could conclude otherwise . . .
. . .
Oracle presented evidence that Java SE was in SavaJe mobile phones and that Oracle licensed Java SE to other smartphone manufacturers, including Danger and Nokia. Because the Java SE was already being used in smartphones, Google did not “transform” the copyrighted material into a new context and no reasonable jury could conclude otherwise."
Bad Faithquote
CAFC also considers bad faith.
"But, while bad faith may weigh against fair use, a copyist’s good faith cannot weigh in favor of fair use. Indeed, the Ninth Circuit has expressly recognized that “the innocent intent of the defendant constitutes no defense to liability. . .
. . .
"Because evidence of Google’s good faith was relevant to rebut evidence of its bad faith, however, and there is no objection to the instructions to the jury on this or any other point, we must assume that the jury simply did not find the evidence of Google’s bad faith persuasive
. . .
"Ultimately, we find that, even assuming the jury was unpersuaded that Google acted in bad faith, the highly commercial and non-transformative nature of the use strongly support the conclusion that the first factor weighs against a finding of fair use."

CAFC concedes to the functional nature of the API, though it ascribes little value to the concession (contrast with the subsequent SCOTUS opinion which gives this factor so much weight that it changes the PNAE order to lead with it). You can tell by the excerpt below that Judge Boudin's Charybdis is at the forefront of CAFC's mind.quote

"Although it is clear that the 37 API packages at issue involved some level of creativity—and no reasonable juror could disagree with that conclusion—reasonable jurors could have concluded that functional considerations were both substantial and important. Based on that assumed factual finding, we conclude that factor two favors a finding of fair use . . . The Ninth Circuit has recognized, however, that this second factor “typically has not been terribly significant in the overall fair use balancing. We note, moreover, that allowing this one factor to dictate a conclusion of fair use in all cases involving copying of software could effectively negate Congress’s express declaration—continuing unchanged for some forty years—that software is copyrightable. Accordingly, though the jury’s assumed view of the nature of the copyrighted work weighs in favor of finding fair use, it has less significance to the overall analysis."

In fairness to the next SCOTUS opinion, though, it's API declaratory code at issue here, not software qua software, and so Congress' intent is a bit more vague.

CAFC ascribes significance to the portion copied, but agrees that it wasn't large relative to all of JAVA.quote
"While, as discussed below, the volume of work copied is relevant to the fair use inquiry generally, thought must be given to the quality and importance of the copied material, not just to its relative quantity vis-à-vis the overall work . . . To hold otherwise would mean that verbatim copying could qualify as fair use as long as the plagiarist stops short of taking the entire work. That approach is inconsistent with settled law and is particularly troubling where, as here, the portion copied is qualitatively significant . . . Even assuming the jury accepted Google’s argument that it copied only a small portion of Java, no reasonable jury could conclude that what was copied was qualitatively insignificant, particularly when the material copied was important to the creation of the Android platform. Google conceded as much when it explained to the jury the importance of the APIs to the developers it wished to attract. See Tr. of Proceedings held on 5/16/16 at 106:8- For these reasons, we find that the third factor is, at best, neutral in the fair use inquiry, and arguably weighs against such a finding."
Unlike SCOTUS, CAFC ascribes considerable importance to this factor.quote
"The Supreme Court once said that factor four is “undoubtedly the single most important element of fair use.” Harper & Row, 471 U.S. at 566. In its subsequent opinion in Campbell, however, the Court emphasized that none of the four factors can be viewed in isolation and that “[a]ll are to be explored, and the results “[m]arket harm is a matter of degree, and the importance of this factor will vary, not only with the amount of harm, but also with the relative strength of the showing on the other factors.” Campbell, 510 U.S. at 590 n.21. On this point, we must apply clear Supreme Court precedent rather than the more recent Ninth Circuit’s statements to the contrary."

So a big issue here is what to do with the 1995 SCOTUS Campbell v. Acuff Rose decision dealing with a parody song. Normally, the Effect on the Market factor was given great weight, but in Campbell, SCOTUS said the factors should be evenly considered.

"The Supreme Court has since clarified that market impact, "no less than the other three [factors], may be addressed only through a 'sensitive balancing of interests'" and that earlier interpretations of Sony to the contrary were incorrect. Campbell, 510 U.S. at 590 n.21 (quoting Sony, 464 U.S. at 455 n.40);see also Monge, 688 F.3d at 1181 (cautioning against overemphasis on a presumption of market harm after Campbell)."

Despite Campbell, the Ninth Circuit has still been giving the fourth factor a lot of weight, but CAFC here does the right thing and says:

"On this point, we must apply clear Supreme Court precedent rather than the more recent Ninth Circuit's statements to the contrary."

CAFC performs the analysis, finding that the record demonstrates considerable market harm.

"[T]he record contained substantial evidence that Android was used as a substitute for Java SE and had a direct market impact. Given this evidence of actual market harm, no reasonable jury could have concluded that there was no market harm to Oracle from Google’s copying. . . .
[A]lthough the district court focused exclusively on the market it found that Oracle had already entered— desktops and laptops—it should have considered how Google’s copying affected potential markets Oracle might enter or derivative works it might create or license others to create. . . That Oracle never built a smartphone device is irrelevant because potential markets include licensing others to develop derivative works . . . a market is a potential market even where the copyright owner has no immediate plans to enter it or is unsuccessful in doing so."

CAFC concludes that the factor weighs heavily against Google, even under the post-Campbell analysis.

"Given the record evidence of actual and potential harm, we conclude that "unrestricted and widespread conduct of the sort engaged in by" Google would result in "a substantially adverse impact on the potential market for the original" and its derivatives. See Campbell, 510 U.S. at 590 (citation and quotation marks omitted). Accordingly, the fourth factor weighs heavily in favor of Oracle."

Summary (N.B., this isn't treated like a "scorecard", the factors may not be given equal weight; still, courts usually phrase their reasoning so a majority of the factors come out on one side; "are to be explored, and the results weighed together, in light of the purposes of copyright.”"):

urpose of the Use: Doesn't favor Fair Use
ature of the copyrighted work: Favors Fair Use (but CAFC accords it less value)
mount of the Work Used: Neutral (or less than neutral)
ffect Upon the Market for the Work: Doesn't Favor Fair Use

CAFC concludes with a summary assessment of Google's activities.

"Although Google could have furthered copyright’s goals of promoting creative expression and innovation by developing its own APIs, or by licensing Oracle’s APIs for use in developing a new platform, it chose to copy Oracle’s creative efforts instead. There is nothing fair about taking a copyrighted work verbatim and using it for the same purpose and function as the original in a competing platform . . . We do not conclude that a fair use defense could never be sustained in an action involving the copying of computer code. Indeed, the Ninth Circuit has made it clear that some such uses can be fair. See Sony, 203 F.3d at 608; Sega, 977 F.2d at 1527-28. We hold that, given the facts relating to the copying at issue here—which differ materially from those at issue in Sony and Sega—Google’s copying and use of this particular code was not fair as a matter of law."

  • This is a good opinion, carefully considering and applying precedent.
  • Arguably the biggest difference (aside from the conclusion :P) between this and the SCOTUS opinion is in the Nature factor analysis. CAFC's analysis is ~750 words, SCOTUS nature analysis is ~1100 words. This is important for SCOTUS, because they will employ the Nature factor in discounting the Effect factor, saying the novel "standards adoption" character of the "nature" of the work, was part of the reason for its "market effect" and therefore the market effect factor should be seen to favor fair use. This results in a somewhat schizophrenic SCOTUS analysis - first saying copying the API didn't cause harm / wasn't a substitutequote, then saying it did cause harm, was a substitute, but that's "not a problem" because of the API's nature as an "adopted" standard (clearly, I'm paraphrasing).quote Personally, I expect it's doubtful Congress anticipated such an analysis in 1976. On the other hand the issue has been present since then, written large in bright red neon since Boudin's concurrence in 1995. At some point qui tacet consentire videtur, and so if lobbyists et al. couldn't be bothered in 44 years to clarify the situation, you can't really blame SCOTUS too harshly for taking the initiative. Still, CAFC's caution in this opinion is to be commended. There may have been no congressional action because the momentum of caselaw seemed to be going in the "desired" direction.
  • Though most news sources / law blogs / etc. haven't celebrated it as much, I think the discussion regarding the jury is almost as important as the Fair Use analysis itself. Even if we weren't entering Boudin's Charybdis, we'd all still benefit from clear reasoning in the precedents. Minimizing the role of the jury will (hopefully) help in that regard. Now that we are entering the Charybdis, though, clear precedents may still not be enough to offset the "host of administrative problems" Judge Boudin feared (recall the inductive ambiguity discussed above in the legal reasoning section - more decisions does not always imply more clarity in the law). Maybe the cost won't be too great, especially if the industry adjusts its expectations. We'll have to see.

Full Opinion

Let's discuss the Majority opinion first and then discuss the Dissent.

Majority Opinion

The Majority Opinion is broken down as follows:

  • Pages 1-11: Procediural and factual background - you've seen this before
  • Pages 12-14: Background on fair use specifically, emphasizing that the statute provides "general principles", which require "judicial balancing, depending upon relevant circumstances, including 'significant changes in technology'".
  • Pages 14-15: Google's petition asked about both copyrightability and fair use. The SCOTUS majority is going to assume ("purely for argument's sake") the former and analyze the latter
  • Pages 15-18: And we're back to background, particularly a policy / philosophical assessment of fair use, explaining why the majority disagrees with the dissent
  • Pages 18-21: The majority deals with Google's arguments for jury determination under the 7th Amendment
  • Pages 21-24: Nature analysis
  • Pages 24-28: Purpose analysis
  • Pages 28-30: Amount used analysis
  • Pages 30-35: Market Effects analysis
  • Pages 35-36: Conclusion and walkback

The philosophy portion on pages 15-18 argues for the policy basis in the decision quote basically laying the groundwork for choosing Boudin's Charybdis:

"The language of §107, the “fair use” provision, reflects its judge-made origins. It is similar to that used by Justice Story in Folsom v. Marsh, 9 F. Cas. 342, 348, F. Cas. No. 4901 (No. 4,901) (CC Mass. 1841). See Campbell, 510 U. S., at 576, 114 S. Ct. 1164, 127 L. Ed. 2d 500 (noting how “Justice Story’s summary [of fair use considerations] is discernable” in §107). That background, as well as modern courts’ use of the doctrine, makes clear that the concept is flexible, that courts must apply it in light of the sometimes conflicting aims of copyright law, and that its application may well vary depending upon context. Thus, copyright’s protection may be stronger where the copyrighted material is fiction, not fact, where it consists of a motion picture rather than a news broadcast, or where it serves an artistic rather than a utilitarian function.
. . .
Generically speaking, computer programs differ from books, films, and many other “literary works” in that such programs almost always serve functional purposes. These and other differences have led at least some judges to complain that “applying copyright law to computer programs is like assembling a jigsaw puzzle whose pieces do not quite fit.” Lotus Development Corp. v. Borland Int’l, Inc., 49 F. 3d 807, 820 (CA1 1995) (Boudin, J., concurring).

I won't belabor this part of the opinion too much - basically the SCOTUS majority is applying "gentle" rhetoric, highlighting points and interpretations amenable to their position, being silent on the others, and inviting Congress to step in if they disagree:

Macaulay once said that the principle of copyright is a “tax on readers for the purpose of giving a bounty to writers.” T. Macaulay, Speeches on Copyright 25 (E. Miller ed. 1913). Congress, weighing advantages and disadvantages, will determine the more specific nature of the tax, its boundaries and conditions, the existence of exceptions and exemptions, all by exercising its own constitutional power to write a copyright statute.

(That quote is actually from the preceding fair use summary on Pages 12-14)

This quote is a great example of the gentle rhetoric applied throughout. "We don't say that, Macaulay said that, but who likes taxes anyway?" Surely less taxes are a good thing. But logically, the assertion is vague. All individual rights are impositions on other folks' rights - does my right not to be murdered "tax" your conduct? I suppose it does, but only certain perspectives would characterize it that way.

Given the rapidly changing technological, economic, and business-related circumstances, we believe we should not answer more than is necessary to resolve the parties’ dispute. We shall assume, but purely for argument’s sake, that the entire Sun Java API falls within the definition of that which can be copyrighted.
Unlike the CAFC opinion, this opinion is saturated with the Nature analysis (which is sort of ironic, since they're supposed to be more evenly considering all the factors under Campbell). This isn't too surprising since it was Nature that Boudin focused upon in his concurrence. In any event, just as the factual posture is rather narrow here, appreciate that the analytic posture is being explicitly made narrow.

Preliminary Jury Issue

This is a short section, basically agreeing with CAFC on the jury issue.quote

We agree with the Federal Circuit’s answer to this question. We have said, “[f]air use is a mixed question of law and fact.” Harper & Row, 471 U. S., at 560. We have explained that a reviewing court should try to break such a question into its separate factual and legal parts, reviewing each according to the appropriate legal standard. But when a question can be reduced no further, we have added that “the standard of review for a mixed question all depends—on whether answering it entails primarily legal or factual work.” U. S. Bank N. A. v. Village at Lakeridge, LLC, 583 U. S. ___, ___, 138 S. Ct. 960, 967, 200 L. Ed. 2d 218 (2018). In this case, the ultimate “fair use” question primarily involves legal work.

Basically, from a historical perspective as regards the 7th amendment objection, this wasn't one of the doctrines in 1791 required to go to a jury, rather, fair use was a matter of equity rather than common law.

Nor is Google correct that “the right of trial by jury” includes the right to have a jury resolve a fair use defense. That Clause is concerned with “the particular trial decision” at issue. Markman v. Westview Instruments, Inc., 517 U. S. 370, 376, 116 S. Ct. 1384, 134 L. Ed. 2d 577 (1996). Even though it is possible to find pre-Revolutionary English cases in which a judge sent related questions like fair abridgment to a jury, those questions were significantly different from the “fair use” doctrine as courts apply it today. See, e.g., Gyles v. Wilcox, 2 Atk. 141, 142-144, 26 Eng. Rep. 489, 490-491 (Ch. 1740) (asking the Court to resolve the narrow question whether a shortened work could be considered a new work); Sayre v. Moore, 1 East 361, n., 102 Eng. Rep. 138, 139, n. (K. B. 1785) (discussing the jury’s role in resolving whether copying constituted infringement). As far as contemporary fair use is concerned, we have described the doctrine as an “equitable,” not a “legal,” doctrine. We have found no case suggesting that application of U. S. Bank here would fail “to preserve the substance of the common-law [jury trial] right as it existed in 1791.” Markman, 517 U. S., at 376, 116 S. Ct. 1384, 134 L. Ed. 2d 577.
Majority Fair Use Analysis

The majority doesn't stick with the normal PNAE order, but I'll reorder their comments below for consistency with the normal approach.

SCOTUS spends most of their time on the "transformative" subfactor, before explaining why they're giving the "commercial" and "bad faith" subfactors less weight. quote

SCOTUS acknowledges that there was literal copying, but "since virtually any unauthorized use of a copyrighted computer program (say, for teaching or research) would do the same, to stop here would severely limit the scope of fair use in the functional context of computer programs."

Here Google’s use of the Sun Java API seeks to create new products. It seeks to expand the use and usefulness of Android-based smartphones. Its new product offers programmers a highly creative and innovative tool for a smartphone environment. To the extent that Google used parts of the Sun Java API to create a new platform that could be readily used by programmers, its use was consistent with that creative “progress” that is the basic constitutional objective of copyright itself. Cf. Feist, 499 U. S., at 349-350, 111 S. Ct. 1282, 113 L. Ed. 2d 358 (“The primary objective of copyright is not to reward the labor of authors, but ‘[t]o promote the Progress of Science and useful Arts’” (quoting U. S. Const., Art. I, §8, cl. 8)).

SCOTUS then references the factual record and amici for support:

The record here demonstrates the numerous ways in which reimplementing an interface can further the development of computer programs. The jury heard that shared interfaces are necessary for different programs to speak to each other. App. 125 (“We have to agree on the APIs so that the application I write to show a movie runs on your device”). It heard that the reimplementation of interfaces is necessary if programmers are to be able to use their acquired skills. Id., at 191 (“If the API labels change, then either the software wouldn’t continue to work anymore or the developer . . . would have to learn a whole new language to be able to use these API labels”). It heard that the reuse of APIs is common in the industry. Id., at 115, 155, 663. It heard that Sun itself had used pre-existing interfaces in creating Java. Id., at 664. And it heard that Sun executives thought that widespread use of the Java programming language, including use on a smartphone platform, would benefit the company. Id., at 130-133.

Amici supporting Google have summarized these same points—points that witnesses explained to the jury. See, e.g., Brief for Copyright Scholars as Amici Curiae 25 (“[T]he portions of Java SE that Google reimplemented may have helped preserve consistency of use within the larger Java developer community”); Brief for Microsoft Corporation as Amicus Curiae 22 (“[A]llowing reasonable fair use of functional code enables innovation that creates new opportunities for the whole market to grow”); Brief for 83 Computer Scientists as Amici Curiae 20 (“Reimplementing interfaces fueled widespread adoption of popular programming languages” (emphasis deleted)); Brief for R Street Institute et al. as Amici Curiae 15-20 (describing Oracle’s reimplementation of other APIs); see also Brief for American Antitrust Institute as Amicus Curiae 7 (“Copyright on largely functional elements of software that [have] become an industry standard gives a copyright holder anti-competitive power”).

The logical import of this litany is not always clear. For example: "Sun executives thought that widespread use of the Java programming language, including use on a smartphone platform, would benefit the company." Surely the assertion isn't that "the purpose of Google's use was to gratify the expectation of Sun executives"? What is this here for then? Presumably it's establishing that some sort of custom or expectation (something like estoppel?) in the community was prevalent and consistent with Google's purpose.

The "commercial" and "bad faith" subfactors are addressed in just two paragraphs:

There are two other considerations that are often taken up under the first factor: commerciality and good faith. The text of §107 includes various noncommercial uses, such as teaching and scholarship, as paradigmatic examples of privileged copying. There is no doubt that a finding that copying was not commercial in nature tips the scales in favor of fair use. But the inverse is not necessarily true, as many common fair uses are indisputably commercial. For instance, the text of §107 includes examples like “news reporting,” which is often done for commercial profit. So even though Google’s use was a commercial endeavor—a fact no party disputed, see 886 F. 3d, at 1197—that is not dispositive of the first factor, particularly in light of the inherently transformative role that the reimplementation played in the new Android system.

As for bad faith, our decision in Campbell expressed some skepticism about whether bad faith has any role in a fair use analysis. 510 U. S., at 585, n. 18, 114 S. Ct. 1164, 127 L. Ed. 2d 500. We find this skepticism justifiable, as “[c]opyright is not a privilege reserved for the well-behaved.” Leval 1126. We have no occasion here to say whether good faith is as a general matter a helpful inquiry. We simply note that given the strength of the other factors pointing toward fair use and the jury finding in Google’s favor on hotly contested evidence, that factbound consideration is not determinative in this context.

These and related facts convince us that the “purpose and character” of Google’s copying was transformative—to the point where this factor too weighs in favor of fair use.

This is probably the most important factor in the analysis (the Majority leads with it). While addressing much Sturm and Drang, the SCOTUS majority basically assesses the API in accordance with Judge Boudin's Charybdis approach, focusing upon interface adoption. quote

The declaring code (inseparable from the programmer’s method calls) embodies a different kind of creativity. Sun Java’s creators, for example, tried to find declaring code names that would prove intuitively easy to remember. Id., at 211. They wanted to attract programmers who would learn the system, help to develop it further, and prove reluctant to use another. See post, at 10 (“Declaring code . . . is user facing. It must be designed and organized in a way that is intuitive and understandable to developers so that they can invoke it”). Sun’s business strategy originally emphasized the importance of using the API to attract programmers. It sought to make the API “open” and “then . . . compete on implementations.” App. 124-125. The testimony at trial was replete with examples of witnesses drawing this critical line between the user-centered declaratory code and the innovative implementing code. Id., at 126-127, 159-160, 163-164, 187, 190-191.

Again, it's a little unclear what role the custom/estoppel type comments imply here.

The majority correctly observes that we're not dealing with your "run of the mill" software here. The following is the core of the whole majority opinion and, really, the core of all of OvG:

These features mean that, as part of a user interface, the declaring code differs to some degree from the mine run of computer programs. Like other computer programs, it is functional in nature. But unlike many other programs, its use is inherently bound together with uncopyrightable ideas (general task division and organization) and new creative expression (Android’s implementing code). Unlike many other programs, its value in significant part derives from the value that those who do not hold copyrights, namely, computer programmers, invest of their own time and effort to learn the API’s system. And unlike many other programs, its value lies in its efforts to encourage programmers to learn and to use that system so that they will use (and continue to use) Sun-related implementing programs that Google did not copy. Although copyrights protect many different kinds of writing, Leval 1116, we have emphasized the need to “recogni[ze] that some works are closer to the core of [copyright] than others,” Campbell, 510 U. S., at 586, 114 S. Ct. 1164, 127 L. Ed. 2d 500. In our view, for the reasons just described, the declaring code is, if copyrightable at all, further than are most computer programs (such as the implementing code) from the core of copyright. That fact diminishes the fear, expressed by both the dissent and the Federal Circuit, that application of “fair use” here would seriously undermine the general copyright protection that Congress provided for computer programs. And it means that this factor, “the nature of the copyrighted work,” points in the direction of fair use.

While "[t]hat fact" may diminish those fears (well, the logical basis for those fears), the majority surely also has Boudin's fear of a "host of administrative problems that would cause cost and delay, and would also reduce the ability of the industry to predict outcomes" in mind. The majority's (arguably Panglossian) celebration of fair use in the policy/philosophy section may be an implicit assertion that they don't consider these problems much of a concern ("The upshot, in our view, is that fair use can play an important role in determining the lawful scope of a computer program copyright, such as the copyright at issue here. It can help to distinguish among technologies").

While arguably just a banal observation about functional character, there's an interesting quote that may be relevant to dynamic linking analysis:

Second, the Sun Java API associates a particular command, called a “method call,” with the calling up of each task. The symbols java.lang., for example, are part of the command that will call up the program (whether written by Sun or, as here, by Google) that instructs the computer to carry out the “larger number” operation. Oracle does not here argue that the use of these commands by programmers itself violates its copyrights.

Though not stated explicitly, the Amount analysis is colored by the Nature analysis.quote

The majority acknowledges that the declaratory code was copied wholesale:

If one considers the declaring code in isolation, the quantitative amount of what Google copied was large. Google copied the declaring code for 37 packages of the Sun JavaAPI, totaling approximately 11,500 lines of code. Those lines of code amount to virtually all the declaring code needed to call up hundreds of different tasks. On the other hand, if one considers the entire set of software material in the Sun Java API, the quantitative amount copied wassmall. The total set of Sun Java API computer code, including implementing code, amounted to 2.86 million lines, of which the copied 11,500 lines were only 0.4 percent. App.212.

However, the majority then looks to the nature of the API to expand its analysis:

Several features of Google’s copying suggest that the better way to look at the numbers is to take into account the several million lines that Google did not copy.
  • For one thing, the Sun Java API is inseparably bound to those task-implementing lines. Its purpose is to call them up.
  • For another, Google copied those lines not because of their creativity, their beauty, or even (in a sense) because of their purpose. It copied them because programmers had already learned to work with the Sun Java API’s system, and it would have been difficult, perhaps prohibitively so, to attract programmers to build its Android smartphone system without them.
  • Further, Google’s basic purpose was to create a different task-related system for a different computing environment (smartphones) and to create a platform—the Android platform—that would help achieve and popularize that objective. The “substantiality” factor will generally weigh in favor of fair use where, as here, the amount of copying was tethered to a valid, and transformative, purpose. Supra, at 25–26; see Campbell, 510 U. S., at 586–587 (explaining that the factor three “enquiry will harken back to the first of the statutory factors, for . . . the extent of permissible copying varies with the purpose and character of the use”).

Eventually, the majority makes explicit that the Nature analysis (as embodied in the Purpose) is dictating its reasoning here:

We do not agree with the Federal Circuit’s conclusion that Google could have achieved its Java-compatibility objective by copying only the 170 lines of code that are “necessary to write in the Java language.” 886 F. 3d, at 1206. In our view, that conclusion views Google’s legitimate objectives too narrowly. Google’s basic objective was not simply to make the Java programming language usable on its Android systems. It was to permit programmers to make use of their knowledge and experience using the Sun Java API when they wrote new programs for smartphones with the Android platform. In principle, Google might have created its own, different system of declaring code. But the jury could have found that its doing so would not have achieved that basic objective. In a sense, the declaring code was the key that it needed to unlock the programmers’ creative energies. And it needed those energies to create and to improve its own innovative Android systems.

We consequently believe that this “substantiality” factor weighs in favor of fair use.

SCOTUS again colors it analysis based upon the Nature factorquote

But a potential loss of revenue is not the whole story. We here must consider not just the amount but also the source of the loss. As we pointed out in Campbell, a “lethal parody, like a scathing theatre review,” may “kil[l] demand for the original.” Id., at 591–592.

Citing parody is a bit odd, as it was the nature of parody that mitigated the market for the copyrighted work. Here, SCOTUS focuses on the nature of the copyrighted work:

On the other hand, Google’s copying helped Google make a vast amount of money from its Android platform. And enforcement of the Sun Java API copyright might give Oracle a significant share of these funds. It is important, however, to consider why and how Oracle might have become entitled to this money. When a new interface, like an API or a spreadsheet program, first comes on the market, it may attract new users because of its expressive qualities, such as a better visual screen or because of its superior functionality. As time passes, however, it may be valuable for a different reason, namely, because users, including programmers, are just used to it. They have already learned how to work with it. See Lotus Development Corp., 49 F. 3d, at 821 (Boudin, J., concurring).

The record here is filled with evidence that this factor accounts for Google’s desire to use the Sun Java API. See, e.g., App. 169-170, 213-214. This source of Android’s profitability has much to do with third parties’ (say, programmers’) investment in Sun Java programs. It has correspondingly less to do with Sun’s investment in creating the Sun Java API. We have no reason to believe that the Copyright Act seeks to protect third parties’ investment in learning how to operate a created work. Cf. Campbell, 510 U. S., at 591-592, 114 S. Ct. 1164, 127 L. Ed. 2d 500 (discussing the need to identify those harms that are “cognizable under the Copyright Act”).

Many would hotly contest that last point. The SCOTUS majority likely has Boudin's discussion of QWERTY in mind here:

"Better typewriter keyboard layouts may exist, but the familiar QWERTY keyboard dominates the market because that is what everyone has learned to use . . . Thus, to assume that computer programs are just one more new means of expression, like a filmed play, may be quite wrong"

Which is a very reasonable observation. The hotly contested issue, however, is: does that difference matter? How do we know that Congress didn't intend to protect interfaces in such situations?

If, as Justice Kennedy observed in his concurrence to Campbell, we should not accord fair use to someone who merely places "the characters from a familiar work in novel or eccentric poses", then why shouldn't we protect an interface applied to a "novel or eccentric" backend? Are we to assume the desire for adoption, without direct remuneration, will itself be sufficient to motivate API designers in the future?

The majority elaborates upon its policy motivations before concluding to discount the factor:

Finally, given programmers’ investment in learning the Sun Java API, to allow enforcement of Oracle’s copyright here would risk harm to the public. Given the costs and difficulties of producing alternative APIs with similar appeal to programmers, allowing enforcement here would make of the Sun Java API’s declaring code a lock limiting the future creativity of new programs. Oracle alone would hold the key. The result could well prove highly profitable to Oracle (or other firms holding a copyright in computer interfaces). But those profits could well flow from creative improvements, new applications, and new uses developed by users who have learned to work with that interface. To that extent, the lock would interfere with, not further, copyright’s basic creativity objectives. See Connectix Corp., 203 F. 3d, at 607; see also Sega Enterprises, 977 F. 2d, at 1523-1524 (“An attempt to monopolize the market by making it impossible for others to compete runs counter to the statutory purpose of promoting creative expression”); Lexmark Int’l, 387 F. 3d, at 544 (noting that where a subsequent user copied a computer program to foster functionality, it was not exploiting the programs “commercial value as a copyrighted work” (emphasis in original)). After all, “copyright supplies the economic incentive to [both] create and disseminate ideas,” Harper & Row, 471 U. S., at 558, 105 S. Ct. 2218, 85 L. Ed. 2d 588, and the reimplementation of a user interface allows creative new computer code to more easily enter the market.

The uncertain nature of Sun’s ability to compete in Android’s market place, the sources of its lost revenue, and the risk of creativity-related harms to the public, when taken together, convince that this fourth factor—market effects—also weighs in favor of fair use.

Note, again, that the court ties its reasoning closely to the facts here at issue:

We do not say that these questions are always relevant to the application of fair use, not even in the world of computer programs. Nor do we say that these questions are the only questions a court might ask. But we do find them relevant here in helping to determine the likely market effects of Google’s reimplementation.

Summary (N.B., this isn't treated like a "scorecard", the factors may not be given equal weight; still, courts usually phrase their reasoning so a majority of the factors come out on one side):

urpose of the Use: Favors Fair Use
ature of the copyrighted work: Favors Fair Use
mount of the Work Used: Favors Fair use
ffect Upon the Market for the Work: Favors Fair Use

Dissenting Opinion

The Dissenting Opinion is broken down as follows:

  • Pages 1-4: Intro and Background
  • Pages 5-9: Argument for Copyrightability
  • Pages 9-11: Nature Analysis
  • Pages 11-14: Market Effect Analysis
  • Pages 15-17: Purpose Analysis
  • Pages 17-19: Amount Analysis

Preliminary Copyrightability Issue

At the outset, note that Thomas doesn't equivocate about copyrightability.

The Court reaches this unlikely result in large part because it bypasses the antecedent question clearly before us: Is the software code at issue here protected by the Copyright Act? The majority purports to assume, without deciding, that the code is protected. But its fair-use analysis is wholly inconsistent with the substantial protection Congress gave to computer code. By skipping over the copyrightability question, the majority disregards half the relevant statutory text and distorts its fair-use analysis. Properly considering that statutory text, Oracle’s code at issue here is copyrightable, and Google’s use of that copyrighted code was anything but fair.

. . .

The Court wrongly sidesteps the principal question that we were asked to answer: Is declaring code protected by copyright? I would hold that it is.

The Dissent takes the more conservative interpretation of the Act.

Computer code occupies a unique space in intellectual property. Copyright law generally protects works of authorship. Patent law generally protects inventions or discoveries. A library of code straddles these two categories. It is highly functional like an invention; yet as a writing, it is also a work of authorship. Faced with something that could fit in either space, Congress chose copyright, and it included declaring code in that protection.

The Copyright Act expressly protects computer code. It recognizes that a “computer program” is protected by copyright. See 17 U. S. C. §§109(b), 117, 506(a). And it 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.” §101. That definition clearly covers declaring code—sets of statements that indirectly perform computer functions by triggering prewritten implementing code.

Again, even though no one's saying it, everyone has Alice v CLS Bank in the back of their mind.

Fair Use Analysis

The dissent first notes oddities in the majority's fair use analysis.

But the majority’s application of fair use is far from ordinary.

By skipping copyrightability, the majority gets the methodology backward, causing the Court to sidestep a key conclusion that ineluctably affects the fair-use analysis: Congress rejected categorical distinctions between declaring and implementing code. But the majority creates just such a distinction. The result of this distorting analysis is an opinion that makes it difficult to imagine any circumstance in which declaring code will remain protected by copyright.

Tellingly, the majority evaluates the factors neither in sequential order nor in order of importance (at least two factors are more important under our precedent) . . . Because the majority’s mistaken analysis rests so heavily on this factor, I begin with it as well.

Though I'm going to use the normal order below . . .

Despite Campbell, the dissent refers to this as the "second most" important factor quote


Regardless, Google fares no better on transformative use. A court generally cannot find fair use unless the copier’s use is transformative. A work is “transformative” if it “adds something new, with a further purpose or different character, altering the first with new expression, meaning, or message.” Campbell, 510 U. S., at 579, 114 S. Ct. 1164, 127 L. Ed. 2d 500. This question is “guided by the examples [of fair use] given in the preamble to §107.” Id., at 578, 114 S. Ct. 1164, 127 L. Ed. 2d 500. Those examples include: “criticism, comment, news reporting, teaching . . ., scholarship, or research.” §107. Although these examples are not exclusive, they are illustrative, and Google’s repurposing of Java code from larger computers to smaller computers resembles none of them. Google did not use Oracle’s code to teach or reverse engineer a system to ensure compatibility. Instead, to “avoid the drudgery in working up something fresh,” id., at 580, 114 S. Ct. 1164, 127 L. Ed. 2d 500, Google used the declaring code for the same exact purpose Oracle did. As the Federal Circuit correctly determined, “[t]here is nothing fair about taking a copyrighted work verbatim and using it for the same purpose and function as the original in a competing platform.” 886 F. 3d, at 1210.

The majority acknowledges that Google used the copied declaring code “for the same reason” Oracle did. Ante, at 1203. So, by turns, the majority transforms the definition of “transformative.” Now, we are told, “transformative” simply means—at least for computer code—a use that will help others “create new products.” Ibid; accord, ante, at 1203 (Google’s copying “can further the development of computer programs”).

That new definition eviscerates copyright. A movie studio that converts a book into a film without permission not only creates a new product (the film) but enables others to “create products”—film reviews, merchandise, YouTube highlight reels, late night television interviews, and the like. Nearly every computer program, once copied, can be used to create new products. Surely the majority would not say that an author can pirate the next version of Microsoft Word simply because he can use it to create new manuscripts.

Ultimately, the majority wrongly conflates transformative use with derivative use. To be transformative, a work must do something fundamentally different from the original. A work that simply serves the same purpose in a new context—which the majority concedes is true here—is derivative, not transformative. Congress made clear that Oracle holds “the exclusive rights . . . to prepare derivative works.” §106(2). Rather than create a transformative product, Google “profit[ed] from exploitation of the copyrighted material without paying the customary price.” Harper, 471 U. S., at 562, 105 S. Ct. 2218, 85 L. Ed. 2d 588.


The second-most important factor—“the purpose and character of the use, including whether such use is of a commercial nature or is for nonprofit educational purposes,” §107(1)—requires us to consider whether use was “commercial” and whether it was “transformative.” Campbell, 510 U. S., at 578-579, 114 S. Ct. 1164, 127 L. Ed. 2d 500. Both aspects heavily favor Oracle.

Begin with the overwhelming commercial nature of Google’s copying. In 2015 alone, the year before the fair-use trial, Google earned $18 billion from Android. That number has no doubt dramatically increased as Android has grown to dominate the global market share. On this scale, Google’s use of Oracle’s declaring code weighs heavily—if not decisively—against fair use.

The majority attempts to dismiss this overwhelming commercial use by noting that commercial use does “not necessarily” weigh against fair use. Ante, at 1204. True enough. Commercial use sometimes can be overcome by use that is sufficiently “transformative.” Campbell, 510 U. S., at 579, 114 S. Ct. 1164, 127 L. Ed. 2d 500. But “we cannot ignore [Google’s] intended purpose of supplanting [Oracle’s] commercially valuable” platform with its own. Harper, 471 U. S., at 562, 105 S. Ct. 2218, 85 L. Ed. 2d 588 (emphasis in original). Even if we could, we have never found fair use for copying that reaches into the tens of billions of dollars and wrecks the copyright holder’s market.

The dissent doesn't discuss Good Faith / Bad Faith.

Though acknowledging the API's highly functional character, the dissent's analysis of this factor is likewise deferential to Congress. quote

This factor requires courts to assess the level of creativity or functionality in the original work. It generally favors fair use when a copyrighted work is more “informational or functional” than “creative.” 4 M. Nimmer & D. Nimmer, Copyright §13.05[A][2][a] (2019). Because code is predominantly functional, this factor will often favor copying when the original work is computer code. But because Congress determined that declaring and implementing code are copyrightable, this factor alone cannot support a finding of fair use.

The majority, however, uses this factor to create a distinction between declaring and implementing code that in effect removes copyright protection from declaring code. It concludes that, unlike implementing code, declaring code is far “from the core of copyright” because it becomes valuable only when third parties (computer programmers) value it and because it is “inherently bound together with uncopyrightable ideas.” Ante, at 1202 - 1203.

Congress, however, rejected this sort of categorical distinction that would make declaring code less worthy of protection.

The dissent calls out the majority on its "inherit binding" rhetoric.

True, declaring code is “inherently bound together with uncopyrightable ideas.” Ante, at1201 - 1203. Is anything not? Books are inherently bound with uncopyrightable ideas—the use of chapters, having a plot, or including dialogue or footnotes. This does not place books far “from the core of copyright.” And implementing code, which the majority concedes is copyrightable, is inherently bound up with “the division of computing tasks” that cannot be copyrighted. Ante, at 1201. We have not discounted a work of authorship simply because it is associated with noncopyrightable ideas. While ideas cannot be copyrighted, expressions of those ideas can. Golan, 565 U. S., at 328, 132 S. Ct. 873, 181 L. Ed. 2d 835.

Finally we get to the key point. The Majority looked at Boudin's Scylla and Charybdis and said: "We're going with the Charybdis."

The dissent looks at Boudin's Scylla and Charybdis and says: "We're not going with either."

Similarly, it makes no difference that the value of declaring code depends on how much time third parties invest in learning it. Many other copyrighted works depend on the same. A Broadway musical script needs actors and singers to invest time learning and rehearsing it. But a theater cannot copy a script—the rights to which are held by a smaller theater—simply because it wants to entice actors to switch theaters and because copying the script is more efficient than requiring the actors to learn a new one.

That's not the analogy I'd use, but the point is still very well said. To the extent there is a choice, it may be more appropriate for Congress than the courts.

The dissent concludes, observing that the Nature factor taints the majority's analysis.

Only after wrongly concluding that the nature of declaring code makes that code generally unworthy of protection does the Court move on to consider the other factors. This opening mistake taints the Court’s entire analysis.

This is a short, three paragraph analysis. quote

Basically, the dissent focuses upon the verbatim copying of the API declaratory code:

Google does not dispute the Federal Circuit’s conclusion that it copied the heart or focal points of Oracle’s work. 886 F. 3d, at 1207. The declaring code is what attracted programmers to the Java platform and why Google was so interested in that code. And Google copied that code “verbatim,” which weighs against fair use. Harper, 471 U. S., at 565, 105 S. Ct. 2218, 85 L. Ed. 2d 588. The majority does not disagree. Instead, it concludes that Google took no more than necessary to create new products. That analysis fails because Google’s use is not transformative. Campbell, 510 U. S., at 586, 114 S. Ct. 1164, 127 L. Ed. 2d 500 (recognizing that this fourth factor “will harken back to the [purpose-and-character] statutory facto[r]”). This factor thus weighs against Google.

Even if Google’s use were transformative, the majority is wrong to conclude that Google copied only a small portion of the original work. The majority points out that the 11,500 lines of declaring code—enough to fill about 600 pages in an appendix, Tr. of Oral Arg. 57—were just a fraction of the code in the Java platform. But the proper denominator is declaring code, not all code. A copied work is quantitatively substantial if it could “serve as a market substitute for the original” work or “potentially licensed derivatives” of that work. Campbell, 510 U. S., at 587, 114 S. Ct. 1164, 127 L. Ed. 2d 500. The declaring code is what attracted programmers. And it is what made Android a “market substitute” for “potentially licensed derivatives” of Oracle’s Java platform. Google’s copying was both qualitatively and quantitatively substantial.

As in various lower court decisions, the dissent considers this the most important factor. quote

By copying Oracle’s code to develop and release Android, Google ruined Oracle’s potential market in at least two ways.

First, Google eliminated the reason manufacturers were willing to pay to install the Java platform. Google’s business model differed from Oracle’s. While Oracle earned revenue by charging device manufacturers to install the Java platform, Google obtained revenue primarily through ad sales. Its strategy was to release Android to device manufacturers for free and then use Android as a vehicle to collect data on consumers and deliver behavioral ads. With a free product available that included much of Oracle’s code (and thus with similar programming potential), device manufacturers no longer saw much reason to pay to embed the Java platform.

For example, before Google released Android, Amazon paid for a license to embed the Java platform in Kindle devices. But after Google released Android, Amazon used the cost-free availability of Android to negotiate a 97.5% discount on its license fee with Oracle. Evidence at trial similarly showed that right after Google released Android, Samsung’s contract with Oracle dropped from $40 million to about $1 million. Google contests none of this except to say that Amazon used a different Java platform, Java Micro Edition instead of Java Standard Edition. That difference is inconsequential because the former was simply a smaller subset of the latter. Google copied code found in both platforms. The majority does not dispute—or even mention—this enormous harm.

The dissent also objects to the majority's speculation upon lock in and public benefit before concluding.

If the majority is going to speculate about what Oracle might do, it at least should consider what Google has done. The majority expresses concern that Oracle might abuse its copyright protection (on outdated Android versions) and “‘attempt to monopolize the market.’” Ante, at 1208 - 1209. But it is Google that recently was fined a record $5 billion for abusing Android to violate antitrust laws. Case AT.40099, Google Android, July 18, 2018 (Eur. Comm’n-Competition); European Comm’n Press Release, Commission Fines Google €4.34 Billion for Illegal Practices Regarding Android Mobile Devices to Strengthen Dominance of Google’s Search Engine, July 18, 2018. Google controls the most widely used mobile operating system in the world. And if companies may now freely copy libraries of declaring code whenever it is more convenient than writing their own, others will likely hesitate to spend the resources Oracle did to create intuitive, well-organized libraries that attract programmers and could compete with Android. If the majority is worried about monopolization, it ought to consider whether Google is the greater threat.

By copying Oracle’s work, Google decimated Oracle’smarket and created a mobile operating system now in over 2.5 billion actively used devices, earning tens of billions of dollars every year. If these effects on Oracle’s potential market favor Google, something is very wrong with our fair-use analysis.

Summary (N.B., this isn't treated like a "scorecard", the factors may not be given equal weight; still, courts usually phrase their reasoning so a majority of the factors come out on one side):

urpose of the Use: Doesn't favor Fair Use
ature of the copyrighted work: "Copyrightable but weakly protected"
mount of the Work Used: Doesn't favor Fair Use
ffect Upon the Market for the Work: Doesn't Favor Fair Use
The dissent concludes:
The majority purports to save for another day the question whether declaring code is copyrightable. The only apparent reason for doing so is because the majority cannot square its fundamentally flawed fair-use analysis with afinding that declaring code is copyrightable. The majorityhas used fair use to eviscerate Congress’ considered policy judgment. I respectfully dissent.

  • Ironically, despite the majority's reliance upon Campbell for "equal consideration of the factors", each of their factor analyses is hard to square with the case law unless the nature factor is allowed to predominate. That's basically what Boudin anticipated in his concurrence for the Charybdis approach, so it's not too surprising.
  • It's odd for ALL the factors to come out one side in a fair use analysis. That the majority's opinion does so speaks to the (not necessarily unjustifiable) infection of the nature analysis in its reasoning.
  • Again, this opinion is very closely tied to the facts at hand (E.g., "The testimony at trial was replete with examples of witnesses drawing this critical line between the user-centered declaratory code and the innovative implementing code.")
  • The majority keeps using the phrase "inextricably boundup with" various things. The point being made appears to be that the API declaratory code serves a referential/functional purpose. The litany includes:
    • "It is inextricably bound together with a general system, the division of computing tasks, that no one claims is a proper subject of copyright."
    • "It is inextricably bound up with the idea of organizing tasks into what we have called cabinets, drawers, and files, an idea that is also not copyrightable."
    • "It is inextricably boundup with the use of specific commands known to programmers, known here as method calls (such as java.lang.Math.max, etc.), that Oracle does not here contest."
    • "And it is inextricably bound up with implementing code, which is copyrightable but was not copied."
    This was probably meant as a rhetorical flourish, dissecting the functional character, which is a fine point to make. However, it goes a bit far - obviously the API wasn't "inextricably bound up" in a literal sense, since Google excised and copied it.
  • Amici citations - As a matter of policy and method, I was very disturbed by the majority's citing of amici in its "purpose and character of the use" factor analysis. Some lawyers may say I'm being petty, but while the citations are ostensibly to "summarize" those amici's assessment of "facts", I would have cautioned against doing so for the following reasons:
    • It doesn't seem necessary - if the goal is to cite or summarize the factual record below, then cite or summarize the factual record below
    • This is a) a very contentious case; b) being decided as a matter of law. Referencing amici in this manner risks an undesirable inference that policy/political motivations of those amici, rather than law, Congress' intentions, and agreed upon facts, are dictating the majority's reasoning. Supreme Court cases should be decided by the Supreme Court.
    • It muddies the analysis. In both of the amici citing paragraphs, the paragraph begins with "The jury heard that . . ." " . . . points that witnesses explained to the jury . . ." but concludes with statements by the amici. The amici were not subject to cross-examination before the jury, indeed, the jury obviously never read these amicus briefs. Form follows function. If these paragraphs' function is to reiterate a factual conclusion of the jury, then reiterate that factual conclusion. If, instead, their function is to express admiration for a particular characterizations of the facts by an amicus on policy grounds, then state the policy characterization explicitly. As it stands, it risks an inference that the majority wanted to do the latter, but only felt comfortable doing so under cover of the former.
  • Bad Faith - As an equitable doctrine, I expected more discussion of bad faith. (Indeed, the Supreme Court referred to fair use as “an equitable rule of reason” in Harper & Row. 471 U.S. at 560). However, aside form the brief reference to Campbell, SCOTUS mostly dismisses the issue.
    As for bad faith, our decision in Campbell expressed some skepticism about whether bad faith has any role in a fair use analysis. 510 U. S., at 585, n. 18, 114 S. Ct. 1164, 127 L. Ed. 2d 500. We find this skepticism justifiable, as “[c]opyright is not a privilege reserved for the well-behaved.” Leval 1126.

    (Leval is a law review article, not a case or statutory authority)

    Note that Harper v. Row, another SCOTUS case, said the opposite: "[Fair use presupposes 'good faith' and 'fair dealing'" which makes sense for a doctrine based in equity. The point here seems to be Campbell's note that "being denied permission to use a work does not weigh against a finding of fair use". That's not the same as saying bad faith doesn't have any role.

Do you think the Supreme Court case was rightly decided?

There are places where the majority's analysis is poorly supported, but as Congress explicitly authorized SCOTUS to make a rolling policy assessment on Fair Use, the opinion is not clearly erroneous.

Personally, given the complexity of the issues, I would have construed Congress' authorization narrowly, either joining with Justice Thomas and Alito in dissent or asking the majority to rewrite the opinion to better ground its reasoning in congressional intent before I signed on.

In a "perfect world" I'd rather SCOTUS err on the side of modest interpretation and have Congress spur them on, rather than err on the side of overreach and be reined in. But I realize it's not that simple, and that I'm certainly not cognizant of all the modern day-to-day interplay between Congress, lobbyists, SCOTUS, etc.

I'm confused. You say SCOTUS decision wasn't erroneous, but I've also heard you say you that the CAFC decision was "better than" the SCOTUS decision. How is that possible? What do you mean?

SCOTUS isn't as bound by precedent as CAFC. So reasoning based upon "policy" isn't as inappropriate at SCOTUS as it would be at CAFC. If they want to add a new shade to a statutory interpretation (e.g., the Fair Use balancing under the Copyright Act), even if it pushes against the weight of the case law, that's their prerogative.

That said, with great power comes great responsibility. Amici are not elected. Amici are not cross-examined before a jury. Amici do not answer before a Senate committee. Every time SCOTUS cites an amici in support of a policy point I wince a little.

(To be clear, I think it's great if amici help courts find propositions, but courts should be very careful not to cite amici as if they were themselves propositions; E.g., "Some of the amici refer to what Google did as 'reimplementation'" - who cares what the amici think? Show me the legal proposition)

When reasoning as a matter of law, the legal propositional base should be thoroughly analyzed.

CAFC did an objectively better job of that. Just consider the below quick/crude relative number of legal citations appearing in each of their analyses:

CAFC 2018
  1. 17 U.S.C. § 107(1).
  2. Wall Data, 447 F.3d at 778
  3. Harper & Row, 471 U.S. at 562
  4. Am. Geophysical Union v. Texaco, Inc., 60 F.3d 913, 921 (2d Cir. 1994);
  5. Infinity Broad. Corp. v. Kirkwood, 150 F.3d 104, 109 (2d Cir. 1998)
  6. Elvis Presley Enters., Inc. v. Passport Video, 349 F.3d 622, 627 (9th Cir. 2003)
  7. A&M Records, Inc. v. Napster, Inc., 239 F.3d 1004, 1015 (9th Cir. 2001)
  8. Campbell, 510 U.S. at 579
  9. Mattel Inc. v. Walking Mountain Prods., 353 F.3d 792, 800
  10. Seltzer, 725 F.3d at 1176
  11. quoting Melville B. Nimmer & David Nimmer
  12. Elvis Presley Enters., 349 F.3d at 629
  13. Worldwide Church of God v. Phila. Church of God, Inc., 227 F.3d 1110, 1117 (9th Cir. 2000)
  14. (quoting Weissmann v. Freeman, 868 F.2d 1313, 1324 (2d Cir. 1989))
  15. Fox News Network, LLC v. TVEyes, Inc., No. 15-3885, 883 F.3d 169, 2018 U.S. App. LEXIS 4786, 2018 WL 1057178
  16. Sony Computer Entertainment, Inc. v. Connectix Corp., 203 F.3d 596 (9th Cir. 2000)
  17. L.A. News Serv. v. CBS Broad., Inc., 305 F.3d 924, 938-39 (9th Cir. 2002) ("Merely plucking the most visually arresting excerpt from LANS's nine minutes of footage cannot be said to have added anything new.").
  18. Folsom v. Marsh, 9 F. Cas. 342, 345, F. Cas. No. 4901 (C.C.D. Mass 1841)
  19. TCA Television Corp. v. McCollum, 839 F.3d 168
  20. Kelly v. Arriba Soft Corp., 336 F.3d 811
  21. Perfect 10, Inc. v. Amazon.com, Inc., 508 F.3d 1146
  22. Fisher, 794 F.2d at 436
  23. Monge, 688 F.3d at 1170
  24. Melville B. Nimmer & David Nimmer, Nimmer on Copyright
SCOTUS 2021 Majority
  1. 17 U. S. C. §107(1).
  2. Campbell., at 579, 114 S. Ct. 1164, 127 L. Ed. 2d 500.
  3. 4 Nimmer on Copyright §13.05[A][1][b] (quoting Netanel, Making Sense of Fair Use, 15 Lewis & Clark L. Rev. 715, 746 (2011)).
  4. Feist, 499 U. S., at 349-350, 111 S. Ct. 1282, 113 L. Ed. 2d 358
  5. U. S. Const., Art. I, §8, cl. 8
SCOTUS 2021 Dissent
  1. §107(1)
  2. Campbell, 510 U. S., at 578-579, 114 S. Ct. 1164, 127 L. Ed. 2d 500.
  3. Harper, 471 U. S., at 562, 105 S. Ct. 2218, 85 L. Ed. 2d 588 (emphasis in original).
CAFC 2018
  1. Campbell, 510 U.S. at 586.
  2. Worldwide Church of God, 227 F.3d at 1118
  3. Harper & Row,
  4. Wall Data, 447 F.3d at 780 (citing Sega Enters. Ltd. v. Accolade, Inc., 977 F.2d 1510, 1519 (9th Cir. 1992)
  5. Feist Publ'ns, Inc. v. Rural Tel. Serv. Co., 499 U.S. 340, 345, 111 S. Ct. 1282, 113 L. Ed. 2d 358 (1991)
  6. Dr. Seuss Enters., L.P. v. Penguin Books USA, Inc., 109 F.3d 1394, 1402 (9th Cir. 1997)
  7. Mattel, 353 F.3d at 803
  8. Fox News Network, 2018 U.S. App. LEXIS 4786, 2018 WL 1057178 (quoting Authors Guild v. Google, Inc., 804 F.3d 202, 220 (2d Cir. 2015)
SCOTUS 2021 Majority
  1. Lotus Development Corp., 49 F. 3d, at 809.
  2. Baker v. Selden, 101 U. S. 99, 25 L. Ed. 841, 1880 Dec. Comm'r Pat. 422 (1880)).
  3. Campbell, 510 U. S., at 586, 114 S. Ct. 1164, 127 L. Ed. 2d 500.
SCOTUS 2021 Dissent
  1. 4 M. Nimmer & D. Nimmer, Copyright §13.05[A][2][a] (2019).
  2. Golan, 565 U. S., at 328, 132 S. Ct. 873, 181 L. Ed. 2d 835.
CAFC 2018
  1. Harper & Row, 471 U.S. at 565. /
  2. Worldwide Church of God, 227 F.3d at 1118
  3. Monge, 688 F.3d at 1179
  4. Seltzer, 725 F.3d at 1178
  5. indirect cites to Kelly v. Arriba Soft Corp., 336 F.3d 811, 820-21 (9th Cir. 2003);
  6. Campbell, 510 U.S. at 586-87
  7. Dr. Seuss Enters., 109 F.3d at 1401
SCOTUS 2021 Majority
  1. 17 U. S. C. §107(4).
  2. Campbell, 510 U. S., at 591-592, 114 S. Ct. 1164, 127 L. Ed. 2d 500
  3. MCA, INC. v. Wilson, 677 F. 2d 180, 183 (CA2 1981)
  4. See 4 Nimmer on Copyright §13.05[A][4]
  5. See Lotus Development Corp., 49 F. 3d, at 821 (Boudin, J., concurring).
  6. Connectix Corp., 203 F. 3d, at 607
  7. Sega Enterprises, 977 F. 2d, at 1523-1524
  8. Lexmark Int’l, 387 F. 3d, at 544
  9. Harper & Row, 471 U. S., at 558, 105 S. Ct. 2218, 85 L. Ed. 2d 588
SCOTUS 2021 Dissent
  1. §107(3).
  2. Harper, 471 U. S., at 565-566, 105 S. Ct. 2218, 85 L. Ed. 2d 588
  3. Campbell, 510 U. S., at 589, 114 S. Ct. 1164, 127 L. Ed. 2d 500.
CAFC 2018
  1. 17 U.S.C. § 107(4)
  2. Harper & Row, 471 U.S. at 566-67.
  3. Campbell, 510 U.S. at 590 (citation and quotation marks omitted).
  4. Infinity Broad., 150 F.3d at 110
  5. Disney Enters., Inc. v. VidAngel, Inc., 869 F.3d 848, 861 (9th Cir. 2017) (citing Leadsinger, 512 F.3d at 531)
  6. Sony Corp. of America v. University City Studios, Inc., 464 U.S. 417, 451, 104 S. Ct. 774, 78 L. Ed. 2d 574 (1984)
  7. A&M Records, 239 F.3d at 1017
  8. Swatch Grp. Mgmt. Servs. Ltd. v. Bloomberg L.P., 756 F.3d 73, 91 (2d Cir. 2014) (citation omitted)
  9. Seltzer, 725 F.3d at 1179
  10. Monge, 688 F.3d at 1182 (quoting Harper & Row, 471 U.S. at 553)
  11. Worldwide Church, 227 F.3d at 1119 (citing Salinger v. Random House, Inc., 811 F.2d 90, 99 (2d Cir. 1987))
  12. Micro Star v. Formgen Inc., 154 F.3d 1107, 1113 (9th Cir. 1998)
  13. Monge, 688 F.3d at 1181
SCOTUS 2021 Majority
  1. 17 U. S. C. §107(4)
  2. Campbell, 510 U. S., at 591-592, 114 S. Ct. 1164, 127 L. Ed. 2d 500
  3. MCA, INC. v. Wilson, 677 F. 2d 180, 183
  4. Nimmer on Copyright §13.05[A][4]
  5. Lotus Development Corp., 49 F. 3d, at 821 (Boudin, J., concurring).
  6. Connectix Corp., 203 F. 3d, at 607;
  7. Sega Enterprises, 977 F. 2d, at 1523-1524
  8. Lexmark Int’l, 387 F. 3d, at 544
  9. Harper & Row, 471 U. S., at 558, 105 S. Ct. 2218, 85 L. Ed. 2d 588
SCOTUS 2021 Dissent
  1. Harper & Row, Publishers Inc., v. Nation Enterprises, 471 U. S. 539, 566, 105 S. Ct. 2218, 85 L. Ed. 2d 588 (1985).
  2. Campbell v. Acuff-Rose Music, Inc., 510 U. S. 569, 592, 114 S. Ct. 1164, 127 L. Ed. 2d 500 (1994).

Obviously there's more to the analysis than just raw number of citations. And again, I expect CAFC partly has more cites because they're aware that they're more bound by precedent. But you can see there's a clear disparity even from this crude side-by-side.

In any event, I had fewer "cringe moments" reading the CAFC opinion, but that's not to say the SCOTUS opinion was "wrong."

Do you think this decision is better for "Big Tech" or for small developers?

I don't think you can categorically say one way or the other. It has pros and cons for each.

That said, I'm concerned that it will be part of a continuing pattern whereby large companies take advantage of markets / ecosystems created by individuals or small groups of individuals.

You create and post interesting content on a social media platform - that platform benefits.

You create and develop an interesting startup - a large company acquires you.

The risk and creative effort was the individuals' - the large companies just wait at the finish line to see who survives. IP (in theory) is supposed to guard against that dynamic, at least providing the individual some negotiating leverage rather than outright capitulation to the larger entity.

Think about programming langauges. Sure, there are sometimes corporate sponsors, but the language designers are often individuals on their own or supported by small supporting foundation: example list.

In contrast, who has the resources to massively rewrite a language's backend for novel hardware? Typically large megacorporations.

So there's a risk here that the pendulum has swung too far the other way.

I've followed your blog / web page for a while. You posed a hypothetical if SCOTUS came out this way. Can a Quenya dictionary be protected post OvG?

This is a very interesting open question.

Quenya is the Elvish language created by J.R.R. Tolkien. The hypo is basically: "If Tolkien writes a Quenya dictionary and someone writes a novel using that dictionary, has OvG made that novel fair use?" There was a Star Trek case with the Klingon language with a similar question that settled before reaching an opinion.

The SCOTUS majority takes pains to explain that it is focusing on software specifically

The fact that computer programs are primarily functional makes it difficult to apply traditional copyright concepts in that technological world. See Lotus Development Corp., 49 F. 3d, at 820 (Boudin, J., concurring). In doing so here, we have not changed the nature of those concepts. We do not overturn or modify our earlier cases involving fair use—cases, for example, that involve “knockoff ” products, journalistic writings, and parodies.
It also gave much significance in its analysis to the "development attraction" aspect of the API:

For one thing, the Sun Java API is inseparably bound to those task-implementing lines. Its purpose is to call them up. For another, Google copied those lines not because of their creativity, their beauty, or even (in a sense) because of their purpose. It copied them because programmers had already learned to work with the Sun Java API’s system, and it would have been difficult, perhaps prohibitively so, to attract programmers to build its Android smartphone system without them.

. . .

Google’s basic objective was not simply to make the Java programming language usable on its Android systems. It was to permit programmers to make use of their knowledge and experience using the Sun Java API when they wrote new programs for smartphones with the Android platform.

With Quenya, arguably it is the "creativity, beauty, purpose" motivating the usage.

So one could distinguish the Quenya hypothetical based upon this difference in factual propositions. You also have several literary cases, protecting characters and other literary elements, that may mitigate the "transformativeness" argument.

Are those differences enough to result in a difference in outcome? I'm not sure, but they're certainly nontrivial distinctions that a court would need to wrestle with.

I'm a cross-compatibility interface developer / open source manager / ecosystem bound start-up / etc. How should I respond to OvG?

It'll depend on your specific facts.

Some folks don't have to do anything. Others should completely abandon their business model.

Many situations can be addressed with creative technical / legal / business strategic hacks. OvG changes the landscape, but you can (often) still reach your destination.