Amicus Curiae Brief — Google LLC, Petitioner v. Oracle America, Inc.

Supreme Court briefJan 10, 2020

Ask Donna

What actually matters in this document.

Text

No. 18-956

================================================================================================================

In The

Supreme Court of the United States

-----------------------------------------------------------------GOOGLE LLC,

Petitioner,

ORACLE AMERICA, INC.,

Respondent.

-----------------------------------------------------------------On Writ Of Certiorari To The

United States Court Of Appeals

For The Federal Circuit

-----------------------------------------------------------------BRIEF OF COPYRIGHT SCHOLARS AS AMICI

CURIAE IN SUPPORT OF THE PETITIONER

-----------------------------------------------------------------REBECCA TUSHNET

Counsel of Record

HARVARD LAW SCHOOL

1575 Massachusetts Ave.

Cambridge, MA 02138

(703) 593-6759

rtushnet@law.harvard.edu

Counsel for Amici Curiae

================================================================================================================

COCKLE LEGAL BRIEFS (800) 225-6964

WWW.COCKLELEGALBRIEFS.COM

TABLE OF CONTENTS

Page

TABLE OF AUTHORITIES .................................

INTEREST OF AMICI CURIAE .........................

SUMMARY OF THE ARGUMENT .....................

ARGUMENT ........................................................

CONCLUSION..................................................... 33

APPENDIX

APPENDIX A: List of Amici Curiae .................... App. 1

TABLE OF AUTHORITIES

Page

CASES

A&M Records, Inc. v. Napster, Inc., 239 F.3d 1004

(9th Cir. 2001).................................................... 12, 13

Alcatel USA, Inc. v. DGI Technologies, Inc., 166

F.3d 772 (5th Cir. 1999) .............................................7

Authors Guild v. Google, Inc., 804 F.3d 202 (2d

Cir. 2015) .................................................................32

Authors Guild, Inc. v. HathiTrust, 755 F.3d 87

(2d Cir. 2014) ...........................................................32

Bateman v. Mnemonics, Inc., 79 F.3d 1532 (11th

Cir. 1996) .................................................................27

Bouchat v. Baltimore Ravens Ltd. Partnership,

737 F.3d 932 (4th Cir. 2013) ....................................16

Cambridge Univ. Press v. Patton, 769 F.3d 1232

(11th Cir. 2014)........................................................30

Campbell v. Acuff–Rose Music, Inc., 510 U.S. 569

(1994) ............................................................... passim

Castle Rock Entertainment Inc. v. Carol Publishing Group, 150 F.3d 132 (2d Cir. 1998) ...................30

Compaq Computer Corp. v. Procom Tech., Inc.,

908 F. Supp. 1409 (S.D. Tex. 1995)............................8

Computer Associates International, Inc. v. Altai,

Inc., 982 F.2d 693 (2d Cir. 1992) ....................... 20, 27

iii

TABLE OF AUTHORITIES—Continued

Page

Experian Information Solutions, Inc. v. Nationwide Marketing Services Inc., 893 F.3d 1176

(9th Cir. 2018)............................................................3

Feist Publ’ns, Inc. v. Rural Tel. Serv. Co., 499 U.S.

340 (1991) ................................................................19

Fisher v. Dees, 794 F.2d 432 (9th Cir. 1986) ...............14

Gulfstream Aerospace Corp. v. Camp Sys. Int’l,

Inc., 428 F. Supp. 2d 1369 (S.D. Ga. 2006) .............16

Harper & Row Publishers, Inc. v. Nation Enters.,

471 U.S. 539 (1985) .................................................15

Lasercomb Am., Inc. v. Reynolds, 911 F.2d 970

(4th Cir. 1990)............................................................8

Lexmark Intern., Inc. v. Static Control Components, Inc., 387 F.3d 522 (6th Cir. 2004) ................... 24

Lotus Development Corp. v. Borland Intern., Inc.,

49 F.3d 807 (1st Cir. 1995) ................................ 11, 27

Matthew Bender & Co. v. West Pub’g Co., 158

F.3d 693 (2d Cir. 1998) ............................................26

Omega S.A. v. Costco Wholesale Corporation,

776 F.3d 692 (9th Cir. 2015) ......................................7

Qad. Inc. v. ALN Associates, Inc., 770 F. Supp.

1261 (N.D. Ill. 1991) ..................................................8

Ross, Brovins & Oehmke, P.C. v. Lexis Nexis

Group, 463 F.3d 478 (6th Cir. 2006) .........................3

Schoolhouse, Inc. v. Anderson, 275 F.3d 726 (8th

Cir. 2002) ...................................................................4

TABLE OF AUTHORITIES—Continued

Page

Sega Enters. Ltd. v. Accolade, Inc., 977 F.2d 1510

(9th Cir. 1992).................................................. passim

Sony Computer Entm’t Inc. v. Connectix Corp.,

203 F.3d 596 (9th Cir. 2000) ............................ passim

STATUTES

17 U.S.C. §102(b) .........................................................20

17 U.S.C. §107 ................................................. 13, 22, 32

OTHER AUTHORITIES

American Committee for Interoperable Systems

(ACIS), Statement of Principles ................................... 16

Mark Bartholomew & John Tehranian, An Intersystemic View of Intellectual Property and

Free Speech, 81 George Washington Law Review 1 (2013) ...........................................................13

Julie E. Cohen et al., Copyright in a Global Information Economy (4th ed. 2015) ...........................5

Ian F. Darwin, Java Cookbook: Solutions and

Examples for Java Developers (3d ed. 2014) .........27

Jay Dratler, Jr., Distilling the Witches’ Brew of

Fair Use in Copyright Law, 43 U. Miami L.

Rev. 233 (1988) ........................................................15

Simon J. Frankel & Matt Kellogg, Bad Faith and

Fair Use, 60 J. Copyright Soc’y U.S. 1 (2012) .........15

TABLE OF AUTHORITIES—Continued

Page

Wendy J. Gordon, How Oracle Erred: The “Use/

Explanation Distinction” and the Future of

Computer Copyright, in Copyright in an Age

of Limitations and Exceptions 319 (Ruth

Okediji, ed., 2017)....................................................24

Justin Hughes, Size Matters (Or Should) in Copyright Law, 74 Fordham L. Rev. 575 (2005)............... 22

Margot E. Kaminski & Guy A. Rub, Copyright’s

Framing Problem, 64 UCLA L. Rev. 1102

(2017) .......................................................................22

Hon. Pierre N. Leval, Toward a Fair Use Standard, 103 Harv. L. Rev. 1105 (1990) ..........................15

Melville B. Nimmer & David Nimmer, Nimmer

on Copyright (2000) ................................................25

Matthew Sag, Predicting Fair Use, 73 Ohio St.

L.J. 47 (2012) ...........................................................13

Pamela Samuelson, The Quest for a Sound Conception of Copyright’s Derivative Work Right,

101 Geo. L.J. 1505 (2013) ........................................33

Alfred Z. Spector, Software, Interface, and Implementation, 30 Jurimetrics J. 79 (1989) .............27

INTEREST OF AMICI CURIAE1

This brief is filed on behalf of the undersigned professors identified in Appendix A.2 Amici are scholars

whose research and teaching focus is copyright law.

Amici have no direct interest in the outcome of this litigation. Amici are concerned that the Court of Appeals

for the Federal Circuit’s decision below misapprehends

the fair use doctrine, and express no opinion on the

first question on which this Court granted certiorari:

whether the declarations that Google used in the Android platform are protectable elements of the Java

Platform Special Edition (Java SE), the work of authorship whose copyright Oracle alleged was infringed.

------------------------------------------------------------------

SUMMARY OF THE ARGUMENT

The fair use doctrine requires courts “to avoid

rigid application of the copyright statute when, on occasion, it would stifle the very creativity which that

law is designed to foster.” Campbell v. Acuff–Rose Music, Inc., 510 U.S. 569, 577 (1994). The Federal Circuit’s

rejection of a jury finding of fair use instead embraced

a rigid approach that, as a matter of law, would bar any

Pursuant to Rule 37.6, Amici affirm that no counsel for a

party authored this brief in whole or in part and that no person

other than Amici and their counsel made a monetary contribution

to its preparation or submission. Petitioner’s consent to the filing

of amicus briefs is filed with the Clerk and Respondent has consented to the filing.

Amici’s institutional affiliations are provided only for purposes of identification.

copying of code into a new program, even for a different

platform, as non-transformative and unfair. To justify

its ruling, the Federal Circuit abandoned a consensus

in the lower courts about the broad scope of fair use

when dealing with highly functional software elements. It made key mistakes about the fair use factors

and their balancing, including inflating the relevance

of commerciality, applying an erroneous “no more than

necessary” standard for copying, dismissing as insignificant the highly functional nature of computer programs, and conflating the market for Java SE as a

whole with the market for individual declarations. Absent these legal errors, it is clear that the jury was at

least reasonable in making factual findings that supported a finding of fair use.

The extent to which a new work has a new meaning, message, or purpose—transformativeness—is often

and rightly prioritized in the fair use analysis. Campbell, 510 U.S. at 579. But what constitutes transformativeness is often contentious. Here, the new purpose of

Google’s new code implementing the declarations was

the creation of a new computing environment in which

Java programmers could readily create programs on

multiple platforms, which required the use of limited

portions of highly functional declarations. This type of

purpose has been recognized as transformative because of its role in furthering competition and innovation. A computer interface supports the creation of

other creative works, and in such situations, it is important to avoid locking in third parties to specific

platforms. See, e.g., Sega Enters. Ltd. v. Accolade, Inc.,

977 F.2d 1510, 1524–27 (9th Cir. 1992).

Factors two and three of the fair use test help define the boundaries of this type of fair use. In cases

such as the one at bar, the highly functional nature of

the copied declarations and the limited amount of the

overall Java SE work used, consistent with industry

practices, are vital considerations supporting the conclusion that Google’s use was a transformative use that

served copyright’s basic goal of encouraging creation of

new works. By discounting to the point of irrelevance

the thinness of protection for highly functional aspects

of computer interfaces (factor two) and of the industry

practice of treating the amount Google reimplemented

as reasonable (factor three), the Federal Circuit distorted its analysis of the other fair use factors, threatening the coherence of fair use doctrine and the

ultimate progress of creativity.

Indeed, because the copyright is thin and Google

copied only a very limited amount, the copying arguably failed to meet the substantial similarity standard.

See, e.g., Experian Information Solutions, Inc. v. Nationwide Marketing Services Inc., 893 F.3d 1176, 1186 (9th

Cir. 2018) (in dealing with thin copyrights, “infringement cannot be based on a showing that only a part of

the work has been copied”); Ross, Brovins & Oehmke,

P.C. v. Lexis Nexis Group, 463 F.3d 478, 482–84 (6th Cir.

2006) (the defendant selected a set of legal forms,

61% of which were forms included in the plaintiff ’s

compilation; for purposes of infringing similarity,

that was not particularly high given the thinness of

plaintiff ’s copyright); Schoolhouse, Inc. v. Anderson,

275 F.3d 726, 729–30 (8th Cir. 2002) (74% overlap

where copyright was thin was insufficient). In the

frame of fair use as well, the Court should not lose

sight of the overriding importance of confining the

scope of rights in highly functional works so as not to

harm the overall creative ecosystem.

By downplaying the relevance of the nature of the

work and the amount taken, the Federal Circuit fell

into the well-known trap of circularity: reasoning that,

because Oracle could have charged a license fee for this

type of use if fair use were unavailable, Oracle suffered

cognizable market harm. Because such claims can be

made for any fair use, which by definition is not paid

for, this reasoning cannot distinguish fair and unfair

uses. But factors two and three can help identify when

crediting such claimed market harm would be inconsistent with copyright’s overall balance between past

and future creators. Thus, given the limited copying of

functional elements here, factors one and four also support fair use, because Google’s purpose was generative

of additional creativity by third parties and because

Oracle’s claimed market harm goes beyond the legitimate scope of its thin copyright in highly functional

declarations.

A thin copyright for software, including Java SE,

provides software copyright owners with meaningful

protection against copying of significant amounts of expression, but meaningful protection does not require

the expansive rights that the Federal Circuit granted.

Providing a broad scope to highly functional elements

of software is unnecessary and dangerous to competition and innovation. Factors two and three enable fair

use to implement this distinction between types of

works. The Federal Circuit erred in not recognizing

this interaction between the fair use factors and instead adopting a rigid rule that would preclude fair use

in computer programs.

------------------------------------------------------------------

ARGUMENT

Oracle’s successful effort before the Federal Circuit to cast the case as a matter of inherently wrongful

copying by a competitor diverts attention away from

the underlying economic principles of copyright. These

principles aim to provide a reasonable degree of protection for copyrightable expression, while leaving

room for others to build upon preexisting works and to

develop new communication technologies.

Factor One: The Jury Could Reasonably Have

Concluded that Google’s Creation of a New

Platform Favored Fair Use.

Factor one, the purpose of the accused use, considers different elements, including transformativeness

and commerciality. This case involves the type of fair

use that one prominent casebook labels “technical interchange,” or enabling the creation of programming options

in software—a specific kind of creativity-promoting

transformativeness. See Julie E. Cohen et al., Copyright

in a Global Information Economy 602 (4th ed. 2015).

Reuse of portions of an API for a functional purpose can be a transformative use of a work (here, Java

SE) where it enables the creation of new creative

works and enhances the overall creative ecosystem.

Copying interfaces to enable programmers to more

easily make programs for multiple platforms is a

transformative purpose because it makes the system

function better for people creating new things. Forcing

needless variation in this context does not enhance

creativity; it balkanizes groups of creators and wastes

effort. By contrast, the jury reasonably could have

found that better cross-platform compatibility allows

programmers to transport their own code to different

implementations; the declarations form building blocks

through which larger structures can be created. The

Federal Circuit thus failed to appreciate the relevance

of factors two and three—the highly functional nature

of the limited number of copied declarations—to a

proper analysis of transformativeness.

a. Promoting Compatibility is Transformative

Because It Promotes Creativity and Innovation in the Larger Software System.

1. The Federal Circuit created a rigid rule

that would preclude fair use in code.

The Federal Circuit reasoned that factor one

weighed against Google because the intrinsic purpose

of the parties’ declarations was the same. Pet. App.

31a-33a. This “intrinsic purpose” reasoning creates a

per se rule against copying any interface elements or

code in a computer program. It fails to weigh the

benefits of standardization for third parties, and fails

to appreciate the interaction of factors two and three

with transformativeness. The Federal Circuit’s simplistic rationale suggests that a law review article that

quoted another law review article for support would

infringe the earlier article, because they both had the

same “intrinsic purpose.” Without assessing the role of

the copied material in the parties’ works, this conclusion is erroneous.

2. Android promoted compatibility for the

Java community, a value distinct from

“interoperability.”

The Federal Circuit accepted Oracle’s argument

that Google’s use was not transformative because it

didn’t achieve complete “interoperability.” That is, any

given Java program could not automatically run on

both mobile and desktop without any changes. Pet.

App. 46a & n.11. But using exact interchangeability to

define the only legitimate purpose for copying declarations or code precludes anyone from using limited

portions of functional libraries without Oracle’s permission—they must license the whole or take nothing.

Oracle’s position would allow it to leverage its rights

over copyrightable aspects of Java SE to forbid legitimate uses of the uncopyrightable aspects. Cf. Omega

S.A. v. Costco Wholesale Corporation, 776 F.3d 692, 693–

94 (9th Cir. 2015) (using copyright in small image to

control uncopyrightable watches was misuse of copyright); Alcatel USA, Inc. v. DGI Technologies, Inc., 166 F.3d

772, 793–94 (5th Cir. 1999) (using copyright to control

uncopyrightable microprocessors was misuse); Lasercomb Am., Inc. v. Reynolds, 911 F.2d 970, 978 (4th Cir.

1990) (leveraging copyright to “suppress any attempt

by the licensee to independently implement the idea

which [software] expresses” was misuse); Qad. Inc. v.

ALN Associates, Inc., 770 F. Supp. 1261, 1269–70 (N.D.

Ill. 1991) (using copyright to control functional and unowned elements of software was misuse).

As other courts have recognized, transformative

compatibility does not require identicality of programs.

The details of the technology and the practices of users

weigh heavily in determining whether the use of aspects of an interface supports new creativity. Sony

Computer Entm’t Inc. v. Connectix Corp., 203 F.3d 596,

607, 609 (9th Cir. 2000) (finding that the goal of compatibility favored fair use; noting that the new console

did not play games as well as the original);3 cf. Compaq

Computer Corp. v. Procom Tech., Inc., 908 F. Supp.

1409, 1419, 1427 (S.D. Tex. 1995) (where defendant

used plaintiff ’s software parameters but chose its own

values for them, producing different outcomes in some

cases, the products were “compatible” for purposes of

false advertising law; “compatibility” is a matter of

practical function). Multiple witnesses testified that

creating a Java programming environment on mobile

devices did not require copying of Java SE as a whole

or even of the APIs as a whole. Indeed, constraints such

See Br. of Connectix Corp., Sony Computer Entertainment

Corp. v. Connectix Corp., No. 99-15852, at 18 (9th Cir. filed May

22, 1999) (Connectix implemented only 137 of the 242 functions

implemented in Sony’s BIOS).

as battery life and the way that multiple applications

would run on a device required avoiding some libraries. JA49-50 (some packages were not appropriate for

mobile because of battery and memory limitations,

among other things); JA264 (wholesale copying of

Java SE would have been inappropriate for mobile);

JA168-70 (offering examples). Google used the libraries that developers would particularly expect and that

would be useful on a mobile platform. JA48-49; JA203;

JA213-14.

The jury could reasonably have concluded that

what Google created served the standards and needs

of the industry—the very reason why interoperability

has value. As Sun’s then-CEO announced to the public after Google launched Android, “Google . . . just

strapped another set of rockets to the [Java] community’s momentum”; he, like others in the industry, considered it to be a Java platform congenial to Java

developers, TX2352, and believed Android would improve the overall health of Java by empowering developers, JA133. In practical terms, the quick success of

Android demonstrated that Oracle’s all-or-nothing

“interoperability” argument did not reflect the perspective of Java programmers, who are the ones who decide

where their own efforts in creating new programs

should be directed.

As a result, the jury could reasonably have agreed

with Google that what is important from the perspective of empowering future creators is sufficient

standardization, much as the Federal Rules of Civil

Procedure standardize key elements of federal practice

despite the existence of local rules. Disallowing programmers from using core elements to which they

have become accustomed makes it much more difficult

for them to work together (or to compete against each

other on the merits of their specific contributions) to

create or adapt innovative apps that can run on both

Android and Java technologies. Importantly, this benefit to the programming community was achieved even

though there are API libraries that are unique to particular environments, such as mobile versus desktop,

and thus not used across all Java programs.

3. Creating a new platform is a purpose favoring fair use because it creates new

works and also enables creation and dissemination of additional works.

The jury could reasonably have found Google’s reimplementation of Java for Android to be a significant

innovation that benefited the larger programming

community. Prior cases have recognized the generative

function of access to platforms. In Sony Computer Entertainment v. Connectix Corp., 203 F.3d 596 (9th Cir.

2000), for example, Connectix reimplemented the functionality of Sony’s interface in its “PlayStation emulator” software. This reimplementation was correctly

held to be transformative for at least two reasons.

First, Connectix’s emulator software enabled consumers to use legitimate, purchased copies of PlayStation

games in a new environment (on personal computers).

Id. at 606–07. Second, the court considered Connectix’s

own creation, the emulator software, “a wholly new

product, notwithstanding the similarity of uses and

functions” between the PlayStation gaming console

and the emulator program. This was at least modestly

transformative because Connectix had created its

own code for implementing the PlayStation firmware’s

functions in the emulator software, furthering the

basic purposes of copyright. Id.

Importantly for purposes of the transformative

use inquiry, both Connectix and Google reimplemented

the functionalities of existing software in new computing environments and wrote their own software code

to implement the relevant interface. Likewise, Sega,

although it was decided before Campbell, recognized

the benefit to creators in general if unlicensed parties

can reimplement interfaces, consistent with Campbell’s recognition of the role of fair use in spurring

new creativity. 977 F.2d at 1522–23 (copying highly

functional elements in order to create a new computing environment was “a legitimate, essentially nonexploitative purpose” which “led to an increase in the

number of independently designed video game programs offered for use with the Genesis console. It is

precisely this growth in creative expression, based on

the dissemination of other creative works and the

unprotected ideas contained in those works, that the

807, 821 (1st Cir. 1995) (Boudin, J., concurring) (suggesting that a commercial use should not infringe

where third parties have invested in their own efforts

that have increased value when they are compatible

across different implementations of software). Although

both Sony and Sega, unlike the Federal Circuit below,

concluded that interfaces (the BIOS in Sony and interface procedures in Sega) were uncopyrightable, their

reasoning on the creativity-promoting transformativeness of the resulting overall software applies with

equal force even assuming that what is incorporated in

the new work is copyrightable.

b. The Federal Circuit inappropriately weighed

commerciality against Google.

Precisely because so many favored fair uses are

for profit, the Court wisely established that the fact

that the use is made by a for-profit endeavor has little

weight against fair use as long as other elements of the

test favor fairness. Campbell, 510 U.S. at 579, 584.

The Federal Circuit went beyond a definition of

commerciality as for-profit use and reasoned that, although Google did not sell its Java implementation,

giving customers for free something they would ordinarily have to buy can constitute commercial use. Pet.

App. 27a (quoting A&M Records, Inc. v. Napster, Inc.,

239 F.3d 1004, 1015 (9th Cir. 2001)). As the Federal Circuit used this formulation, it is mistaken. First, it is

factually erroneous: customers never have to buy

Java—Oracle demanded that Google, but not ordinary

programmers, purchase a license.

Second and more significantly, the framing of commerciality as a matter of whether payment is ordinary is both inconsistent with the plain meaning of

“commercial”4 and mistakenly duplicates factor four

(the effect on the market for the original work). See

Sony, 464 U.S. at 450 n.33 (rejecting the argument that

a use is necessarily commercial if it replaces a sale);

Mark Bartholomew & John Tehranian, An Intersystemic View of Intellectual Property and Free Speech,

81 George Washington Law Review 1, 22 (2013)

(“Courts have stretched the definition of commercial

use to include activities that ordinarily would not be

viewed as commercial in any other context, especially

when First Amendment rights are at issue. . . . [Market harm] concerns are already covered by the fourth

factor of the test and need not be redundantly considered in the first.”); Matthew Sag, Predicting Fair Use,

73 Ohio St. L.J. 47, 60 (2012) (“[T]his construction of

commerciality is almost exactly equivalent to the concept of market effect under the fourth factor, so employing it here would be double counting to say the

least.”).5

Webster’s relevantly defined the adjective “commercial” as

“engaged in work designed for the market,” “of or relating to commerce,” “characteristic of commerce,” “viewed with regard to

profit,” and “designed for a large market,” with “commerce” relevantly defined as “the exchange or buying and selling of commodities on a large scale. . . .” Webster’s New Collegiate Dictionary

226 (1973).

The Federal Circuit’s statement that “[d]irect economic

benefit is not required to demonstrate a commercial use,” 886

F.3d at 1198 (quoting A&M, 239 F.3d at 1015), is also suspect,

as the Federal Circuit applied it. “[N]onprofit educational uses,”

17 U.S.C. §107(1), may contribute to the quality of education,

leading to additional donations or resources for the institution;

successful criticism may enhance a scholar’s career. Given the

c. Good or bad faith are generally unhelpful

terms in fair use, but, to the extent it is considered, Google acted in good faith.

In its arguments below, Oracle argued that

Google’s unsuccessful pursuit of a license for additional rights to trademarks and copyrights evinced bad

faith, which should weigh against Google in the factor

one analysis. Pet. App. 95a. Oracle’s reasoning flies in

the face of this Court’s sound precedent. Campbell, 510

U.S. at 585 n.18 (“2 Live Crew’s actions [in unsuccessfully seeking permission] do not necessarily suggest

that they believed their version was not fair use; the

offer may simply have been made in a good-faith effort

to avoid this litigation. If the use is otherwise fair, then

no permission need be sought or granted.”) (citing

Fisher v. Dees, 794 F.2d 432, 437 (9th Cir. 1986) (“[T]o

consider Dees blameworthy because he asked permission would penalize him for this modest show of consideration.”)).

As the district court noted, the license about which

the parties had been negotiating was for a wider use of

the full range of the Java technologies (as well as for

co-branding). Pet. App. 107a. Even if such extensive

uses required a license, Google could reasonably believe that using only some of the Java declarations and

classes was lawful.

breadth with which the Federal Circuit, and some other courts,

have applied the concept of indirect benefit, no fair use could escape condemnation for “commerciality,” demonstrating this formulation’s lack of utility as a consideration in a balancing test.

More generally, amici believe that “good faith” is

often an unhelpful framing in factor one. See Campbell, 510 U.S. at 585 n.18; Hon. Pierre N. Leval, Toward a Fair Use Standard, 103 Harv. L. Rev. 1105,

1128 (1990) (“Whether the secondary use is within

the protection of the [fair use] doctrine depends on

factors pertinent to the objectives of the copyright

law and not on the morality or motives of either the

secondary user or the copyright-owning plaintiff.”). It

has little relationship to the statute’s language or

purpose:

[F]rom the standpoint of faithfulness to statutory language, a user’s course of dealing with

the holder of copyright in the underlying work

has little relation to the ‘purpose’ of the use.

Second, and more important, there is little

reason to infuse the doctrine of fair use with

notions of commercial ethics. Unlike the doctrine of trade secrecy, the doctrine of fair use

has no substantial basis in commercial morality. Like copyright law generally, fair use has

an economic purpose.

Jay Dratler, Jr., Distilling the Witches’ Brew of Fair

Use in Copyright Law, 43 U. Miami L. Rev. 233, 334

(1988) (footnotes omitted); see also Simon J. Frankel

& Matt Kellogg, Bad Faith and Fair Use, 60 J. Copyright Soc’y U.S. 1, 36 (2012) (reaching the same conclusions).6 However, as discussed below, in cases

The Court in Harper & Row Publishers, Inc. v. Nation Enters., 471 U.S. 539 (1985), characterized good faith as part of fair

use. There, however, the defendant preempted the copyright

owner’s well-recognized right of first publication, which it had

involving only limited copying, a justified belief that

the copying was appropriately limited bears on the

reasonableness of the amount taken and thus on the

third factor.

To the extent that good or bad faith matters independently, the Federal Circuit’s treatment was illogical: it stated that only bad faith was relevant, and that

evidence of good faith could only be used to rebut bad

faith. Pet. App. 39a. The jury could legitimately take

into account Google’s reasonable belief in the rightness

of its acts to support fair use. See, e.g., Bouchat v. Baltimore Ravens Ltd. Partnership, 737 F.3d 932, 942 (4th

Cir. 2013); Gulfstream Aerospace Corp. v. Camp Sys.

Int’l, Inc., 428 F. Supp. 2d 1369, 1378 n.5 (S.D. Ga.

2006).

Among other things, the American Committee for

Interoperable Systems (ACIS), of which both Sun Microsystems and Oracle were members, endorsed and

helped entrench the general understanding that reimplementing APIs was not infringing. ACIS said: “The

rules or specifications according to which data must be

organized in order to communicate with another program or computer, i.e., interfaces and access protocols,

are not protectable expression under copyright law.”

See American Committee for Interoperable Systems

(ACIS), Statement of Principles, available at https://www.

ccianet.org/wp-content/uploads/2014/10/ACIS-Letter-toClinton-Admin-1992.pdf. Sun was a founding member

planned to exploit. Bad and good faith are neither necessary for

such a result nor helpful outside that context.

of ACIS, and Oracle joined ACIS in 1992. Brief Amicus

Curiae of American Committee for Interoperable Systems at iii, Eng’g Dynamics, Inc. v. Structural Software,

Inc., 46 F.3d 408 (5th Cir. 1995) (No. 92-3444) (listing

ACIS members). ACIS also told this Court:

Unlike traditional literary works such as novels and plays that stand alone and do not need

to interact with any other work, computer programs never function alone; they function

only by interacting with the computer environment in which their developers place

them. . . . As a consequence, no matter how

much better or cheaper the new program is,

it will not enjoy a single sale if it cannot interoperate in its intended environment. If the

developer of one part of the environment can

use copyright law to prevent other developers

from writing programs that conform to the

system of rules governing interaction within

the environment—interface specifications, in

computer parlance—the first developer could

gain a patent-like monopoly over the system

without ever subjecting it to the rigorous scrutiny of a patent examination.

Brief Amici Curiae of American Committee for Interoperable Systems and Computer & Communications Industry Ass’n in Support of Respondent, Lotus

Dev. Corp. v. Borland Int’l, Inc., 516 U.S. 233 (1996) (No.

94-2003), 1995 WL 728487 at *4-5.

Factor Two: The Federal Circuit Erred by Discounting the Highly Functional Nature of the

Work and the Thinness of the Copyright

The “nature” of the work includes both the quantity of originality and the role of that originality in how

the work is used or received by the public. The jury

heard evidence that the declarations and classes of the

Java SE API were functional, not merely in the way

that all computer code performs a function, but specifically in that these particular declarations perform

their mini-duties in noncreative ways. Pet. App. 113a14a. The Federal Circuit stated that the jury could reasonably have found that the Java API as a whole merited only “thin” copyright protection. Pet. App. 42a.7

The Federal Circuit correctly labeled the nature of

the work as favoring Google, but it failed to put this

factor in proper context and thus to give it actual

weight, instead reasoning that the second factor did

not matter to fair use determinations generally. Pet.

App. 42a-43a. Although the fictional nature of a work

may not weigh against fair use when an accused use

is transformative in a way that creates a new, critical

meaning or message, Campbell, 510 U.S. at 586, that

does not mean that the second factor should be discounted in every case. There are four factors because

various fair uses can be fair for different reasons, and

the factors may thus differ in relevance across types of

cases. Unsurprisingly, none of the cases cited by the

As stated above, Amici take no position on the copyrightability of the API as such.

Federal Circuit in support of always discounting the

second factor dealt with functional aspects of computer programs. Cases that have considered “technical

interchange” situations, such as Connectix, 203 F.3d at

603, have wisely placed special emphasis on the highly

functional features of interfaces.

Here, the reasons that factor two favored Google

are highly relevant to the overall fair use analysis. The

Ninth Circuit has observed:

The second statutory [fair use] factor . . . reflects the fact that not all copyrighted works

are entitled to the same level of protection. . . .

Works of fiction receive greater protection

than works that have strong factual elements,

such as historical or biographical works, or

works that have strong functional elements,

such as accounting textbooks.

Sega, 977 F.2d at 1524 (citations omitted). Copyright

protection “does not extend to the ideas underlying a

work or to the functional or factual aspects of the

work.” Id. As a result, even if copyright protects a

highly functional work as a whole, “[u]nder the Copyright Act, if a work is largely functional, it receives only

weak protection.” Id. at 1527. This result, said the

court, was “neither unfair nor unfortunate. It is the

means by which copyright advances the progress of science and art.” Id. (quoting Feist Publ’ns, Inc. v. Rural

Tel. Serv. Co., 499 U.S. 340, 350 (1991)).

The functional character of computer programs

means that they “hover” near “the elusive boundary

line described in [17 U.S.C.] §102(b).” Computer Associates International, Inc. v. Altai, Inc., 982 F.2d 693, 704

(2d Cir. 1992). Owing to software’s hybrid role as both

“literary expression” and a “highly functional, utilitarian component in the larger process of computing,”

this narrow scope “flows from applying, in accordance

with Congressional intent, long-standing principles of

and how many choices went into producing Java SE,

the highly utilitarian nature of declarations means

that copyright grants them thin scope at best.

This thinner scope of protection naturally leads to

a broader scope for fair use, especially in conjunction

with factor three. Connectix, 203 F.3d at 603, 605 (fair

use was an appropriate way to “preserve[ ] public access to the ideas and functional elements embodied

in copyrighted computer software,” and factor two

“strongly favored” fair use); Sega, 977 F.2d at 1522,

1526 (same). The functional nature of software favors

fair use because of the desirability of enabling secondcomers to build on the functional elements of existing

programs in creating new works of authorship. In the

software context, then, finding fair use when there is

partial rather than wholesale copying is particularly

likely to serve copyright’s purpose of encouraging new

creativity.

Factor Three: The Reasonable Amount Taken

from the Accusing Work as a Whole, Java SE

Google took only a small amount—the amount it

deemed appropriate for Java programmers to be able

to program for Android—which strongly favored fair

use. As this Court has explained, the extent of permissible copying varies with the purpose and character of

the use. Campbell, 510 U.S. at 586–87. The question is

not whether, in judicial hindsight, the defendant took

more than absolutely necessary, but whether the

amount taken was “reasonable in relation to the purpose of the copying.” Id. at 586. The Federal Circuit disregarded this governing standard by applying a far

more restrictive and unpredictable rule: considering

the amount taken to be at best neutral and “arguably”

to weigh against Google, Pet. App. 47a, because it was

“more than necessary” to write in Java, Pet. App. 45a.

The Federal Circuit’s rule also ignored the context

of software interfaces. Because any copyright in highly

functional elements is necessarily thin, limited copying should weigh significantly in favor of fair use.

Without such a rule, thin copyrights will expand past

their statutory boundaries. Moreover, as the jury heard

and evidently credited, Google took an amount from

Java SE considered in the industry to be reasonable in

light of its purpose of developing new, compatible

works. Given the limited copying, this factor strongly

favored Google. Sega, 977 F.2d at 1526–27.

1. Google took a small part of Java SE, the

relevant work.

One aspect of the Federal Circuit’s error was its

emphasis on the 37 packages that Google copied in “entirety.” Pet. App. 45a. In a fair use analysis, it is vitally

important to identify the allegedly infringed work or

works. As the statute commands, the proper inquiry

considers the amount taken “in relation to the copyrighted work as a whole.” 17 U.S.C. §107. Without careful attention to the boundaries of works, especially

unfamiliar works such as software, plaintiffs can manipulate their claims to artificially increase the relative size of what was taken. Justin Hughes, Size

Matters (Or Should) in Copyright Law, 74 Fordham L.

Rev. 575, 579–80 (2005) (explaining dangers of recognizing “microworks” in software and elsewhere); id. at

613 (“If our goal is to create special incentives for the

building of houses, we do not necessarily need special

incentives for the making of bricks or the mixing of

mortar. . . . Without such evidence that our bricks—

short phrases, titles, evaluations—are under baked, so

to speak, we should prevent the law from moving in

that direction. . . .”); Margot E. Kaminski & Guy A.

Rub, Copyright’s Framing Problem, 64 UCLA L. Rev.

1102, 1118–19 (2017) (explaining that the Federal Circuit’s earlier decision “did not explain why the quantitative component of its inquiry should zoom in on this

one small function rather than conceive of it as a subpart of the much larger software”); id. at 1142–44 (explaining the interaction between work size and fair

use; noting the risks of a plaintiff “gaming” the work’s

size).

Here, the registrations as well as industry practice

make clear that the work at issue is Java SE.8 All of

Java SE had about 5 million lines of code, including

aspects such as the Java Virtual Machine and a Java

compiler. JA212.9 Google’s implementation expressing

the declarations at issue (the ones that Google reimplemented in its own code from the 37 packages at issue) totals about 11,500 lines, out of what was, at the

2008 release of Android, up to 15 million lines of code.

Id. Even within the subcategory of the 37 packages,

Google’s own implementations accounted for 97% of

the code. Pet. App. 219a. Google needed only those

parts of the Java SE API that were appropriate for the

smartphone context (as compared with the laptop and

server systems for which Java was originally designed). As a quantitative matter, what was copied

was a rounding error—for both Android and Java SE.

Yet the Federal Circuit characterized Google as “taking

a copyrighted work verbatim,” Pet. App. 53a, as if the

fractional part were the whole.

Oracle’s registrations submitted to the court were for Java

SE 1.4 (for “new and revised computer code and accompanying

documentation and manuals,” TX464), and the same for Java SE

5.0, TX475, along with a supplemental certification of registration

with alternate titles for Java SE 5.0, TX476.

The part of Java SE that was the libraries (the full 166

packages) had about 2.86 million lines of code. JA212.

2. Copying highly functional elements weighs

strongly in favor of fair use.

The jury reasonably could have found that the

qualitative character of the copying—its limitation

to highly functional declarations—likewise militates

strongly in favor of fair use. The jury reasonably could

have accepted that Google was indifferent to the expression of Java SE and sought only the function. See

Wendy J. Gordon, How Oracle Erred: The “Use/Explanation Distinction” and the Future of Computer Copyright, in Copyright in an Age of Limitations and

Exceptions 319, 320 (Ruth Okediji, ed., 2017) (“[A] copyright owner should have no prima facie rights over

copying behavior where (1) the goals of the copying are

‘use’ (behavior in the realm of utility patent) and (2)

the copying is done solely for goals unrelated to the expressiveness of the plaintiff ’s work of authorship. . . .

Interoperability is one of the few areas where indifference to expression is clear: After all, when one wants a

spare key made, the elegance or beauty of the key’s

shape is irrelevant—all that matters is that the shape

fits the lock.”). Cf. Lexmark Intern., Inc. v. Static Control Components, Inc., 387 F.3d 522, 544 (6th Cir. 2004)

(if a programmer designed a program that would only

work with the input of a copyrightable poem as the key,

then copyright imposes no barrier to use of the key,

even if the poem would be protected outside that context).

Along with ignoring the highly functional nature

of what was taken in its qualitative evaluation, the Federal Circuit engaged in erroneously circular reasoning.

By reasoning that, because copying benefited Google in

making Android, there was qualitative significance,

Pet. App. 46a, the Federal Circuit remade factor three

so that it would always weigh against fair use. See 1

Melville B. Nimmer & David Nimmer, Nimmer on Copyright §13.05[A][3], at 13–180 (2000) (“To avoid circular reasoning, the plaintiff manifestly should not be

heard to argue that the defendant’s copying of brief

passages vouchsafes their qualitative significance.”

(footnote omitted)).

As the district court detailed, the jury could have

found that Google used highly limited portions of Java

SE’s APIs as part of an innovative mobile device platform. Pet. App. 117a. Google not only wrote its own implementing code for the portions of Java SE that it

used, but also created many new declarations to enable

a vast array of additional, innovative smartphone

functionalities. Id. at 105a. Furthermore, the portions

of Java SE that Google reimplemented may have

helped preserve consistency of use within the larger

Java developer community. Id. at 104a-05a. Given this,

the jury could reasonably have concluded that Google’s

use of limited portions of Java SE was appropriate to

achieve its purpose.

Oracle also argued that Google copied the structure, sequence and organization (SSO) of the Java SE

API libraries. Because of the ways in which declarations are named, that SSO necessarily came with the

names, but the jury could reasonably have found that

this was not relevantly “additional” copying of creative

expression, any more than the pagination of a print

volume of caselaw that naturally follows from it having

cases in some order adds additional expression. See

Matthew Bender & Co. v. West Pub’g Co., 158 F.3d 693,

699 (2d Cir. 1998). If it is fair to use the individual declarations, then it is fair to use the SSO that necessarily

comes along with them.

Testimony in the fair use trial record, including by

Sun’s last CEO,10 supported Google’s contention that

there was a common understanding in the software industry that programmers were free to reimplement

declarations in independently written code.11 This reasonable belief bears on the appropriateness of the

quantitative and qualitative amount taken by Google

in writing its own code. As the District Court noted,

Google’s witnesses

testified that they had understood that “reimplementing” an API library was a legitimate, recognized practice so long as all that

was duplicated was the “declaring code” and

so long as the duplicator supplied its own “implementing code,” that is, the methods were

JA113-15, 119-20 (testifying to an understanding from the

outset of Java’s release that implementations were protected but

API itself was not).

See, e.g., JA246 (testifying to “established industry practice” that the API could be reimplemented without permission);

JA164-65 (same); JA153, 157 (same; witness was employed at

Sun during the relevant period); JA126-28, 55 (openness of the

API “was a way to try to make things fair”; that free use of APIs,

even by competitors, was acceptable and promoted accessibility of

Java to programmers; that APIs were “open” and “then we compete on implementations” without being “locked into one company’s implementation”).

“re-implemented.” In this way, Java programmers using the Android API could call on functionalities with the same Java command

statements needed to call the same functionalities in the Java API, thereby avoiding

splintering of the ways that identical functionalities became invoked by Java programmers.

Pet. App. 96a. The distinction between interfaces and

implementations is long-standing in the computing

field. See, e.g., Alfred Z. Spector, Software, Interface,

and Implementation, 30 Jurimetrics J. 79, 85–86

(1989). The industry consensus was strengthened by

several appellate court cases, including two significant Ninth Circuit precedents, that rejected copyright

claims involving computer program interfaces.12

See, e.g., Sega, 977 F.2d at 1514–15, 1522 (characterizing

the Sega interface reimplemented by unauthorized game developer

as unprotectable “functional requirements for compatibility”);

Connectix, 203 F.3d at 602–03 (software that emulated a game

platform’s functionality by reimplementing its BIOS interface

in independently written code, allowing owners of PlayStation

games to play on unlicensed consoles, used only unprotectable

material); see also Altai, 982 F.2d at 714–15 (rejecting infringement claim based on copying the structure, sequence, and organization of plaintiff ’s list of services as well as parameter lists that

set forth interfaces for interacting with third-party programs because the list of services “was dictated by the nature of other programs with which it was designed to interact” and the parameter

lists were not similar enough to infringe); Bateman v. Mnemonics,

Inc., 79 F.3d 1532, 1539, 1543–46 (11th Cir. 1996) (holding that

jury must be allowed to consider evidence and argument that

some literal copying of code was necessary so that its new operating system could execute its own application program); Lotus, 49

Indeed, a large number of books have been published that set forth the Java SE API, in whole or in

part, including the declarations and class structures.

See, e.g., Ian F. Darwin, Java Cookbook: Solutions

and Examples for Java Developers (3d ed. 2014). These

books aim to explain how to use the Java API to develop new software programs. These for-profit publications often reproduce the whole or substantial parts

of the Java API. The District Court took note of these

books in explaining its denial of Oracle’s Rule 50 motion:

Many thousands of pre-written methods have

been written for Java, so many that thick

books . . . are needed to explain them, organized by packages, classes, and methods. For

each method, the book sets forth the precise

declaring code but does not (and need not) set

forth any implementing code. In other words,

the book duplicates all of the method declarations (organized by packages and classes) together with plain English explanations. A

Java user can study the book and learn the

exact method name and inputs needed to invoke a method for use in his or her own program. . . .

Pet. App. 101a-02a. Where others were freely copying

large parts or the entirety of the API, the jury could

F.3d at 815–18 (holding that the selection and arrangement of

specific commands for invoking specific functions of Lotus’s popular 1-2-3 spreadsheet program was unprotectable).

have found that Google’s far more limited copying

strongly favored Google.

In this way, the thinness of the copyright—factor

two—interacts with factor three: even if APIs as a

whole cross the line into copyrightability, Google

should be able to use declarations (and the SSO they

necessarily reflected) that were reasonably necessary

to pursue its legitimate goal of enabling the creation of

an environment accessible to Java programmers.

The Federal Circuit dismissed Google’s reasons for

copying by claiming that “there is no inherent right to

copy in order to capitalize on the popularity of the

copyrighted work or to meet the expectations of intended customers.” Pet. App. 46a. Not only is this reasoning wrongly indifferent to the limited qualitative

and quantitative amount copied, it disregards the importance of competition and third-party creation. The

work at issue is not a popular film that can stand on

its own, but a platform for creating new programs.

While there is no right to copy wholesale simply because a work has achieved marketplace success, its

dominance can justify taking reasonable amounts in

order to create a new platform that enhances the capabilities of third parties to create new works.

Given the testimony about the common understanding about freedom to reimplement APIs, bolstered by precedent and by commercial copying of the

Java API in books, the amount taken was appropriately limited based on Google’s purpose and strongly

favored fair use.

Factor Four: Circularity and the Federal Circuit’s Conflation of the Market for Java SE

with the Hypothetical Market for Declarations

The Federal Circuit’s analysis of factor four was

entirely and improperly circular: If what Google did

was fair use, then Oracle would lose a potential license

fee. But that is always true when fair use is found. Pet.

App. 268a (“[F]air use asks whether the accused infringer needed a license for its use in the first place.

Thus, consideration of harm to the potential market to

license works like accused infringer’s begs the question

of fair use.”).

In some situations involving traditional creative

works with “thick” copyrights, experience and normative commitments to free speech break the circle: the

markets for criticism, educational uses, and parody

are not legitimate even if a specific copyright owner

evinces a willingness to license particular instances.

Campbell, 510 U.S. at 592 (“[T]here is no protectible

derivative market for criticism.”); Castle Rock Entertainment Inc. v. Carol Publishing Group, 150 F.3d 132,

146 n.11 (2d Cir. 1998) (holding that a copyright owner

cannot control fair use markets merely “by developing

or licensing a market for parody, news reporting, educational or other transformative uses of its own creative work”); see also Cambridge Univ. Press v. Patton,

769 F.3d 1232, 1278 (11th Cir. 2014) (“Plaintiffs may

not head off a defense of fair use by complaining that

every potential licensing opportunity represents a potential market for purposes of the fourth fair use factor.”).

When it comes to software, the benefits of interoperability and allowing third parties to avoid lock-in to

particular platforms similarly explain why licensing

interface reimplementation in newly written code is

likewise not a legitimate copyright market. In Sega, for

example, blocking access to a larger market was not a

legitimate economic interest under factor four, even

though doing so would benefit Sega economically and

even though Sega had established a licensing market

for its interface. 977 F.2d at 1523–24.

Here, as in Sega and Connectix, factors two and

three can be used to avoid the vice of circularity. See

Hughes, supra, at 619 (discussing interaction between

small amount of copying and market effect). The jury

could reasonably have found that there is a fundamental difference between, on the one hand, full copies of

Java SE or revisions that alter Java SE as a whole sufficiently to constitute a new derivative work, and, on

the other, limited copying of highly functional aspects

of Java SE.

In reversing the jury’s verdict, the Federal Circuit

wrongly relied on the circular claim that Google could

have taken a license for the precise use it made and

that this possibility established market harm. Pet.

App. 51a. But the evidence did not establish harm to

the market for “the work,” §107(4), or even the existence of a market for a limited number of highly functional declarations. As Larry Page testified, the parties’

licensing negotiations were for far broader uses of

Java SE, and included proposed trademark/brand licensing. JA245. The Federal Circuit, in substituting its

view of the evidence for the jury’s, ignored the obvious

differences between the market for licensing all aspects of Java SE and the market for a small number of

declarations within specific libraries. Those two things

are not the same, any more than the market for an entire book is the same as the market for short critical or

illustrative excerpts. This distinction is especially salient because, to the extent that Oracle seeks the right

to license functional aspects of its APIs, it attempts to

extend its rights beyond that which copyright allows;

this cannot be considered a traditional, reasonable, or

likely to be developed market for the rights conferred

by copyright. Authors Guild v. Google, Inc., 804 F.3d

202, 224 (2d Cir. 2015) (cognizable market effect must

be based on copyrightable aspects of what was copied,

not on uncopyrightable aspects); Connectix, 203 F.3d at

607–08 (competition is not market harm); Sega, 977

F.2d at 1523–24 (same).

Thus, the jury’s fair use verdict did not undermine Oracle’s derivative work right; by its plain

terms, Section 107 limits all of the copyright owner’s

exclusive rights, including the derivative work right.

17 U.S.C. §107; Authors Guild, Inc. v. HathiTrust, 755

F.3d 87, 95 (2d Cir. 2014). Oracle argued that Android

harmed the market for Java ME, its own mobile product, but the jury heard testimony from Sun’s former

head of ME licensing that Java ME was created before

Java SE and contained no identifiable expression from

the work in suit, Tr. 1668:13-19, Tr. 1691:22-1692:11;

competition with a different product is not cognizable

market harm.

The derivative work right has an important role to

play, but that role is not to shrink fair use. See Pamela

Samuelson, The Quest for a Sound Conception of

1505, 1538 (2013). When what is taken from the copyright claimant’s work is subject to “thin” protection at

best, as discussed under Factor Two, fair use assures

that the derivative work right does not vitiate the welljustified limits on the scope of a copyright owner’s

rights in a highly factual or functional work.

------------------------------------------------------------------

CONCLUSION

Because the Federal Circuit ignored the special

characteristics of software interfaces and applied the

fair use factors rigidly, its fair use analysis was distorted. It gave highly functional software protection

far beyond what is necessary to prevent piracy, to the

detriment of competition and creativity generally. Its

fair use holding should be reversed.

Respectfully submitted,

REBECCA TUSHNET

Counsel of Record

HARVARD LAW SCHOOL

1575 Massachusetts Ave.

Cambridge, MA 02138

(703) 593-6759

rtushnet@law.harvard.edu

Counsel for Amici Curiae

This is a copy of a public record, reproduced as it was published. It is not legal advice, and it may not be the version a court would rely on. Check the official source before you cite it.