Oracle America, Inc. v. Google Inc.

Filing 1138

RESPONSE to re #1116 Trial Brief Oracle's May 14, 2012 Copyright Liability Reply Brief by Oracle America, Inc.. (Attachments: #1 Exhibit A)(Jacobs, Michael) (Filed on 5/14/2012)

1
2
3
4
5
6
7
8
9
10
11
12
13
14
15
16
17

MORRISON & FOERSTER LLP
MICHAEL A. JACOBS (Bar No. 111664)
[e-mail]
KENNETH A. KUWAYTI (Bar No. 145384)
[e-mail]
MARC DAVID PETERS (Bar No. 211725)
[e-mail]
DANIEL P. MUINO (Bar No. 209624)
[e-mail]
755 Page Mill Road, Palo Alto, CA 94304-1018
Telephone: [phone] / Facsimile: [phone]
BOIES, SCHILLER & FLEXNER LLP
DAVID BOIES (Admitted Pro Hac Vice)
[e-mail]
333 Main Street, Armonk, NY 10504
Telephone: [phone] / Facsimile: [phone]
STEVEN C. HOLTZMAN (Bar No. 144177)
[e-mail]
1999 Harrison St., Suite 900, Oakland, CA 94612
Telephone: [phone] / Facsimile: [phone]
ORACLE CORPORATION
DORIAN DALEY (Bar No. 129049)
[e-mail]
DEBORAH K. MILLER (Bar No. 95527)
[e-mail]
MATTHEW M. SARBORARIA (Bar No. 211600)
[e-mail]
500 Oracle Parkway, Redwood City, CA 94065
Telephone: [phone] / Facsimile: [phone]
Attorneys for Plaintiff
ORACLE AMERICA, INC.

18
19

UNITED STATES DISTRICT COURT

20

NORTHERN DISTRICT OF CALIFORNIA

21

SAN FRANCISCO DIVISION

22

ORACLE AMERICA, INC.

23

Plaintiff,

24

v.

25

Case No. CV 10-03561 WHA
ORACLE’S MAY 14, 2012
COPYRIGHT LIABILITY REPLY
BRIEF

GOOGLE INC.

26

Defendant.

27
28
ORACLE’S MAY 14, 2012 COPYRIGHT LIABILITY REPLY BRIEF
CASE NO. CV 10-03561 WHA
pa-1528434

Dept.: Courtroom 8, 19th Floor
Judge: Honorable William H. Alsup

1
2
3
4

Question 1: Copyrightability of Programming Languages
Copyright can protect an original combination of vocabulary and grammar in a computer
programming language. Google’s brief misstates both U.S. and European law.
Remarkably, Google claims “the ‘structure’ or ‘organization’ of words in relation to other

5

words is not protectable.” ECF No. 1116 at 2. If Google were correct, the detailed structure of

6

novels and other writings could be freely copied. Google cites no authority and this is not the

7

law. See, e.g., Urantia Foundation v. Maaherra, 114 F.3d 955, 959 (9th Cir. 1997) (selection,

8

arrangement and structure of answers to questions from purportedly divine aliens copyrightable).

9

Google also now claims the interdepencies in the Java APIs are “nothing special” citing only to

10

the highly simplified analogy Dr. Reinhold used. But at trial its witnesses attested to the

11

creativity and skill API design requires. See RT at 736:25-752:14 (Bloch); 2209:7-8 (Astrachan).

12

Google next contends, again without authority, that it could only be liable for copying the

13

SSO “as a whole.” ECF No. 1116 at 3. This position is inconsistent with the jury instruction

14

Google proposed at the charging conference, which the Court essentially adopted. RT at

15

2401:19-2403:20. It is also wrong under the law. “[A] copyright defendant need not copy a

16

plaintiff's work in its entirety to infringe that work.” L.A. Printex Indus., Inc. v. Aeropostale, Inc.,

17

2012 U.S. App. LEXIS 7079 at * 22 (9th Cir. Apr. 9, 2012) (citations omitted).

18

In addition, Google claims “nothing in the record suggests that Google copied the

19

sequence in which the APIs are implemented within the source code.” ECF No. 1116 at 2.

20

Actually, Google’s own expert so testified. See RT at 2214: 3-9 (“Q. And the Structure,

21

Sequence and Organization of the API elements is virtually identical across those 37 packages,

22

correct? / A. That’s right.”). Against this straightforward admission, Google offers an example of

23

one class where the order of methods in Android source code differs from Java. ECF No. 1116 at

24

2-3. This proves nothing. The order in which the methods appear within each class is not what

25

matters from a technical or structural standpoint. What matters is that the methods appear within

26

the same classes and packages as in the 37 Java APIs and have the same relationships to other

27

elements. The sequence of parameters in a method’s parameter list also matters and Google

28

copied thousands of those. RT at 2212:7-13 (Astrachan). Moreover, the record shows Google
ORACLE’S MAY 14, 2012 COPYRIGHT LIABILITY REPLY BRIEF
CASE NO. CV 10-03561 WHA
pa-1528434

1

1

did copy the order of methods in the source code. See, e.g., ECF No. 1115 at 2 (citing source

2

code). This is another example of Google copying more than it required for “compatibility.”1

3

Google mistakenly claims the CJEU held in SAS that “The programming language is not a

4

‘form of expression of that program’ for purposes of copyright law.” ECF No. 1116 at 4. The

5

CJEU held only that programming languages are not a form of expression of a computer program

6

for purposes of protecting computer programs under the Software Directive. Case C-406/10, SAS

7

Inst., Inc. v. World Programming Ltd., Judgment (May 2, 2012) ¶ 39. It ruled that programming

8

languages (and data formats) could be copyrightable under the Copyright Directive if they

9

constitute the “author’s own intellectual creation.” SAS ¶ 45. SAS also states that copying the

10

“choice, sequence and combination” of keywords, syntax and commands from a user manual

11

could constitute infringement under the Copyright Directive. Id. ¶¶ 66-67, 71.

12

Google also misquotes SAS as saying protecting programming languages would “amount

13

to making it possible to monopolise ideas, to the detriment of technological progress and

14

industrial development.” ECF No. 1116 at 4-5 (quoting SAS ¶ 40). The sentence actually refers

15

to protecting “the functionality of a computer program” only. See SAS ¶ 40.

16

Lastly, SAS did not concern “in essence, the question in the present case.” ECF No 1116

17

at 4. Less copyrightable expression was at issue. In SAS the defendant did not copy “any of the

18

structural design of the source code.” SAS ¶ 25. A jury has found against Google on that point.

19

In addition, SAS did not concern APIs. See Case C-406/10, SAS Inst., Inc. v. World Programming

20

Ltd., (Nov. 29, 2011) ¶ 79 (file formats were “blank forms” with read/write locations).

21

Question 2: Google Copied More Than Names and Headers of Declarations

22

Google admits it “used the same method declarations.” ECF No. 1116 at 5. This means it

23

copied the names, parameters, parameter sequence, return types, and “throws” clauses (exception

24

lists) in the Java APIs. See id. at 5. Google did not have to copy most of these elements to design

25

its own method with similar functionality. To use the example of the password authentication

26
27
28

1

Methods are not just ordered “alphabetically” in the documentation. ECF No. 1116 at 3.
In both Java and Android, the shorter method summaries are in alphabetical order. But in Java
the more extensive “method detail” section follows the order in the source code. Android is
inconsistent. In any event, the complex API organization is not just alphabetical.
ORACLE’S MAY 14, 2012 COPYRIGHT LIABILITY REPLY BRIEF
CASE NO. CV 10-03561 WHA
pa-1528434

2

1

from Oracle’s opening brief (ECF No. 1118 at 13), Google could have rearranged the order of the

2

eight parameters, changed their names, given the method a different name or chosen to throw an

3

exception, all without changing the underlying idea of obtaining a password given certain inputs.

4

By copying the method names and declarations, Google copied Oracle’s design and

5

structure. The methods are a creative and essential part of the structure that represent years of

6

design effort. Unrebutted evidence showed Sun and Oracle had many possible method choices in

7

designing these 37 API packages. RT at 597:9-13, 627:21-629:5; 630:11-631:18 (Reinhold); id.

8

at 1238:11-1240:8; 1240:23-1244:16 (Mitchell). Many creative decisions are involved in

9

something as simple as choosing the methods for drawing a rectangle. ECF No. 1118 at 5-6. By

10

copying the SSO of thousands of these methods, Google infringed. No engineering team

11

independently could have come up with all these same solutions. RT at 1249:18-25 (Mitchell).

12

Google’s argument that “the declarations are, in essence, the titles of the things they

13

declare” is incorrect. ECF No. 1116 at 5. The method declarations in the Java APIs are not titles

14

or names. They express essential parts of the SSO and are copyrightable as such. In Applied

15

Innovations Inc. v. Regents of Univ. of Minn., for example, the court rejected the argument that

16

short, declarative sentences in the plaintiff’s psychology test, such as “I am a good mixer,” were

17

uncopyrightable because of their significance in the context of the test. 876 F.2d 626, 634-35 (8th

18

Cir. 1989) (“The test statements are short, simple, declarative sentences, but they are not merely

19

fragmentary words and phrases within the meaning of 37 C.F.R. § 202.1(a). They are not names

20

or titles or slogans.”). In Health Grades, Inc. v. Robert Wood Johnson Univ. Hosp., Inc., the

21

court found the plaintiff stated a copyright infringement claim based on copying of its ratings and

22

awards such as “five stars” and “clinical excellence” because they represented sufficiently

23

original compilations of fact. 634 F. Supp. 2d 1226, 1238 (D. Colo. 2009). The court noted that

24

there is a “paucity of cases holding that an otherwise original expression is uncopyrightable solely

25

because it is a short phrase” and concluded that the copyright regulation on this topic, 37 C.F.R.

26

202.1(a) “does not strip copyright protection from such original expression.” Id. (citation

27

omitted). Accord Compaq Computer Corporation v. Ergonome, Inc., 137 F. Supp. 2d 768, 774 -

28

775 (S.D. Tex. 2001). See also CCC Info. Servs., Inc. v. Maclean Hunter Mkt. Reports, Inc.,
ORACLE’S MAY 14, 2012 COPYRIGHT LIABILITY REPLY BRIEF
CASE NO. CV 10-03561 WHA
pa-1528434

3

1
2

44 F.3d 61, 67 (2d Cir. 1994) (upholding copyrightability of used car pricing guide).
Similarly, in West Publ’g Co. v. Mead Data Central, Inc., 799 F.2d 1219, 1227 (8th Cir.

3

1986), the court found that West was likely to prevail on its claim that LEXIS copied its

4

pagination. The court emphasized that in copying the pagination, the defendant was copying

5

West’s arrangement of decisions: “The key to this case, then, is not whether numbers are

6

copyrightable, but whether the copyright on the books as a whole is infringed by the unauthorized

7

appropriation of these particular numbers.” Id. In Matthew Bender & Co. v. West Publ’g Co., the

8

Second Circuit reached the opposite conclusion on the same facts, but for reasons that are clearly

9

distinguishable here: the pagination was automatically generated and did not reflect “even a

10

modicum of creativity” and West's arrangement was not copied 158 F.3d 693, 699-700 (2d Cir.

11

1998). Here creativity and Google’s copying of the SSO were both undisputed.

12

Question 3: Fully Qualified Names Can Have More Than Three Parts

13

In stating that the fully qualified name “has three parts,” (ECF No. 1116 at 6), Google

14

overlooks subpackages and nested Member classes and interfaces. See ECF No 1118 at 7.

15

Question 4: Google Could Have Come Up With Its Own Names and SSO

16

Google admits that “It would have been possible in many instances for Google to have

17

created APIs with different names and/or SSO that would have provided similar functionality.”

18

ECF No. 1116 at 6. Its legal and factual arguments for why it did not do so are meritless.

19

The elements of the 37 API packages that are required to implement the Java language are

20

minimal. See Questions 6-8 infra. Google’s primary argument is that industry custom and

21

developer demand required it to copy these 37 API packages. But Google cannot use the demand

22

for Oracle’s copyrighted product as an “external factor” constraining Google’s development.

23

None of the cases it cites support its position; all but one have been addressed previously. The

24

only new case it cites for this point, Swirsky v. Carey, shows just how far Google has strayed: the

25

scenes a faire issue in that case was whether the plaintiffs’ chorus was commonplace because it

26

bore resemblance to a third party melody, the folk song, “For He’s a Jolly Good Fellow.”

27

376 F.3d 841, 850 (9th Cir. 2004). Necessity was not at issue.

28
ORACLE’S MAY 14, 2012 COPYRIGHT LIABILITY REPLY BRIEF
CASE NO. CV 10-03561 WHA
pa-1528434

4

1

Google’s factual arguments are equally flawed. It argues that without any APIs, the Java

2

language is unusable. But when Sun released Java in 1996 there were only 7 packages, including

3

just 4 of the 37 at issue here. Yet Java immediately became enormously popular and could be

4

used to accomplish many useful things. RT at 631:19-25; 686:6-7 (Reinhold); TX 980.

5

Even if developers demand more, Google could have designed its own APIs.

6

Dr. Astrachan’s rote testimony on why developers would demand all 37 packages was shaky at

7

best. For example, he claimed developers would expect the database packages java.sql and

8

javax.sql based on little more than that he would not know how to write the code himself. See

9

RT at 2199:7-13 (Astrachan). Although the Court requested it, Google neglected to ask

10

Dr. Astrachan whether Google could have designed these APIs itself. See id. at 2197:25-2198:23.

11

When Oracle asked, Dr. Astrachan admitted it. RT at 2213:5-19 (Astrachan).

12

The Court should disregard Google’s belated attempt to introduce additional evidence

13

about Spring. Mr. Ellison was Oracle’s first witness. Dr. Astrachan was permitted to respond to

14

issues that arose at trial and could have responded to this one. See RT at 2195:10-25. But if the

15

Court does look to the Spring website, it shows Mr. Ellison’s testimony was correct as to Java

16

Enterprise Edition. Spring is an enterprise application development framework, and Springsource

17

created a new environment, including dozens of new APIs with similar functionality to the Java

18

EE APIs. It had to train a whole new group of developers on those APIs. RT at 304:16-22

19

(Ellison). Oracle also proved that third parties created alternatives to a number of the Java SE

20

APIs, including the 37 at issue. See Oracle May 10 Brief, ECF No. 1118 at 8.

21

Question 5: A Combination of Input-Output Schemes May Be Protectable

22

Neither party contends the input-output scheme for an individual method alone is

23

copyrightable. The original combination of all such elements in the Java APIs, however, is

24

copyrightable. Engineering Dynamics, cited by Google, supports Oracle. It recognized the

25

selection of input/output formats for a computer program may be copyrightable if sufficiently

26

original. Eng’g Dynamics, Inc. v. Structural Software, Inc., 46 F.3d 408, 410 (5th Cir. 1995).

27

Moreover, like the other cases Google cites, it views the merger doctrine from the plaintiff’s

28

perspective: “Consequently, as our opinion explains, the district court will inquire on remand
ORACLE’S MAY 14, 2012 COPYRIGHT LIABILITY REPLY BRIEF
CASE NO. CV 10-03561 WHA
pa-1528434

5

1

whether EDI exercised any judgment in formulating the input cards or merely reflected the

2

industry standards and laws of engineering.” Id. (emphasis added). The Court placed a specific

3

burden on Google prove merger and scenes a faire. ECF No 433 at 9. Google never met it with

4

respect to the combination of input-output schemes or any other elements of the APIs.

5

Questions 6-8: Relationship Between APIs and the Java Programming Language

6

Google urges that all 37 packages are “core” to the Java programming language. ECF

7

No. 1116 at 11-12. No witness testified to this definition, and Google concedes that the

8

documentation for J2SE 5.0, the version relevant here, does not use the term “core” to refer to any

9

subset of API packages. ECF No. 1116 at 11 n.7.

10

The term “core” should not be used as a basis for the Court’s decision. Rather than

11

speculate based on marketing descriptions on a book cover or loose terminology, the Court should

12

look to the Java Language Specification (3rd ed.) (“JLS”), which both parties’ witnesses agreed

13

defines exactly what the language requires. TX 984; RT at 641:4-642:25 (Reinhold); RT at

14

776:3-778:9, 780:24-781:1 (Bloch). The parties also agree that the language directly references

15

only 60 or 61 classes. TX 1062; RT at 676:1-681:2 (Reinhold); RT at 777:21-24 (Bloch).

16

The JLS shows that for most classes the language merely requires the existence of a class

17

with a particular name. TX 1062; RT at 676:14-678:13 (Reinhold) (“There is no mention of what

18

methods might be in them, what fields they might have. They could have anything, as far as the

19

language specification is concerned.”); id. at 681:22-682:2. Google could therefore have

20

designed entirely different classes as long as they had these names. For other classes, the JLS

21

requires the presence of only one or a handful of methods or fields. Again, Google could have

22

designed entirely different classes as long as they had these methods and fields. Exhibit A is a

23

complete list of the page numbers in TX 984 where the Court can find these methods and fields.

24

Google vastly overstates the number of references that these 61 classes make to elements

25

in other packages, or that are made by the 3 packages java.lang, java.io and java.util. See ECF

26

No. 1118 at 11. But this is irrelevant in any event. The JLS does not require any of them. The

27

JLS does not require that any identified class, method, or field to have any relationship

28

whatsoever with any API element in any other package. .
ORACLE’S MAY 14, 2012 COPYRIGHT LIABILITY REPLY BRIEF
CASE NO. CV 10-03561 WHA
pa-1528434

6

1

Finally, Google misconstrues TX 1063, which is simply a list of the 39 classes and

2

interfaces that the compiler mentions, not what the JLS requires. RT at 679:22-681:21. No

3

witness claimed at trial that TX 1063 should be used as the list of classes the language requires.

4

Question 9: Interdependencies in the Implementation

5

The parties’ experts agreed that the SSO of Google’s implementation of these 37 API

6

packages is virtually identical to Oracle’s. ECF No. 1118 at 12. When Google says the

7

implementations are different, it is using the word “implementation” to refer to the method bodies

8

alone, ignoring that structural identity. The parties agree that there was no testimony at the trial

9

on the extent to which interdependencies in the method bodies of Java and Android are similar.

10
11

Question 10: Interdependencies in Names and Declarations
Google’s brief contains an incomplete list of the interdependencies that exist at the name

12

and declaration level. The Court should look to Oracle’s brief. See ECF No. 1118 at 12-13.

13

Question 11: Google Misconstrues the Holding In the ADA Case

14

After repeatedly touting Judge Easterbrook’s copyright expertise (see, e.g., ECF No. 778

15

at 3), Google now asks the Court to disregard his opinion in ADA because it is “nonsensical” and

16

“poorly reasoned.” ECF No. 1116 at 13, 15. Google claims the case “has only been cited three

17

times by any court.” ECF No. 1116 at 13 (emphasis in original). This is false. Shepardizing

18

ADA shows more than 25 decisions around the country have cited it.

19

Google also argues ADA does not actually hold a taxonomy can be copyrighted and claims

20

the Seventh Circuit has cited the decision “only once, and only for propositions unrelated to the

21

‘taxonomy’ holding.” ECF No. 1116 at 13. Google is wrong again. In Edgenet, Inc. v. Home

22

Depot U.S.A., Inc., 658 F.3d 662, 666 (7th Cir. 2011), Judge Easterbrook described his opinion in

23

ADA as “holding that taxonomies are copyrightable.” The ATC case Google cites questioned only

24

whether a numbering system alone could be copyrightable, and agreed with ADA that

25

“Classification schemes can in principle be creative enough to satisfy the originality requirement

26

of copyright protection.” ATC Distrib. Grp., Inc. v. Whatever It Takes Trans. & Parts, Inc.,

27

402 F.3d 700, 706 (6th Cir. 2005) (citing ADA). See also Kendall Holdings, Ltd. v. Eden

28

Cryogenics LLC, 2012 U.S. Dist. LEXIS 5245, at *18-19 (S.D. Ohio Jan. 17, 2012) (quoting
ORACLE’S MAY 14, 2012 COPYRIGHT LIABILITY REPLY BRIEF
CASE NO. CV 10-03561 WHA
pa-1528434

7

1
2

same ADA language with approval).
Google next argues “nothing in the decision supports the conclusion” that the ADA’s

3

numbering system was copyrightable. ECF No. 1116 at 13-14. This is wrong too. ADA states

4

“even the short description and the number are original works of authorship.” 126 F.3d at 979

5

(emphasis added). It explains why the numbering is copyrightable and then again states that “all

6

three elements of the Code – numbers, short descriptions, and long descriptions, are copyrightable

7

subject matter.” Id. Other cases recognize ADA held the numbering system alone was

8

copyrightable. See, e.g., Southco, Inc. v. Kanebridge Corp., 258 F.3d 148, 155 (3d Cir. 2001).

9

Google argues that the taxonomy in ADA “is an unprotectable system.” ECF No. 1116

10

at 14. But the copyrightability of taxonomies is well accepted. Google still has never explained

11

what it means by “system.” As ADA states, “A dictionary cannot be called a ‘system’ just

12

because new novels are written using words, all of which appear in the dictionary. Nor is word-

13

processing software a ‘system’ just because it has a command structure for producing

14

paragraphs.” ADA, 126 F.3d at 980. Google tries to distinguish the APIs claiming that, unlike

15

the numbering system in ADA, the APIs come with “instructions for use” like “a recipe for a new

16

dish” and are the means by which developers express themselves. ECF No. 1116 at 14 (quoting

17

ADA 126 F.3d at 980). It is not clear what difference this would make or what the reasoning was

18

for the Seventh Circuit’s reference to “instructions for use.” Like the Code taxonomy at issue in

19

ADA, APIs describe the various elements in the libraries and the relationships among them.

20

While the APIs can be used as a blueprint to implement the class libraries, Google’s witnesses

21

repeatedly stated at trial that the APIs are not the equivalent of a recipe for developers to write

22

Java programs. See, e.g., RT at 769:4-12 (Bloch). And even if they were a “system,” the detailed

23

expression of that system, which Google copied, is still protectable.

24

The ADA code was copyrightable even though the ADA encourages the code’s use by

25

insurers, dentists and others. ADA, 126 F.3d at 981. Like the ADA, Oracle does not contend

26

developers cannot invoke the Java APIs in their applications. Its complaint is Google creating an

27

unlicensed, incompatible copy of the Java APIs. This was the same reason the ADA sued,

28

because the defendant “used most of the code but made modifications.” See id. The court found
ORACLE’S MAY 14, 2012 COPYRIGHT LIABILITY REPLY BRIEF
CASE NO. CV 10-03561 WHA
pa-1528434

8

1

it could not “make and distribute a derivative work based on the Code.” Id. Instead of furthering

2

compatibility, Google chose to “embrace, extend and extinguish,” copying enough of the Java

3

APIs to capture Java developers and move them over to its own incompatible platform.

4

Question 12: Google Misconstrues CDN v. Kapes

5

The parties agree Kapes treated the plaintiff’s price estimates as a compilation. ECF

6

No. 1116 at 15. But Google is wrong in stating the court determined “that the coin prices in its

7

guide were not of its own creation.” See id. (citing CDN Inc. v. Kapes, 197 F.3d 1256, 1260 (9th

8

Cir. 1999)). The Ninth Circuit agreed with the district court “that the prices in CDN’s guides are

9

not facts, they are ‘wholly the product of [CDN’s] creativity.’” CDN, 197 F.3d at 1260.

10

The API packages are also entitled to protection under the Kapes definition of

11

compilation. Oracle’s quarrel with the statutory definition is that it states a compilation is based

12

on “the collection and assembling of preexisting materials or of data.” 17 U.S.C. § 101. See ECF

13

No. 853 at 1. The APIs are also original works of authorship. Oracle is not claiming registration

14

as a collective work, but has consistently claimed protection for the combination of elements

15

contained in the 37 API packages. See, e.g., ECF No. 339 at 15-16 (SJ Opp.), ECF No. 853 at 1

16

(Apr. 3, 2012 brief), ECF No. 997 at 4-5 (comments on jury instructions). This principle is

17

recognized in cases such as Satava v. Lowry, 323 F.3d 805, 810-11 (9th Cir. 2003) (considering

18

combination of unprotectable elements in sculpture that was original work of authorship) and

19

Merch. Transactions, 2009 U.S. Dist. LEXIS 25663 at *46. The 37 API packages are protectable

20

under the principles expressed in those cases, regardless of how they were registered or labeled.

21

Question 13: Google’s Definition of Compatibility Is Incorrect

22

Google states, without citing any authority, that what matters in this case is whether

23

Android “is compatible with the APIs in the 37 packages in the computer science sense,” by

24

which its means that code “written using the APIs in [the 37] packages will work on both

25

platforms.” ECF No. 1116 at 16. This is not the proper definition. See, e.g., Creative Labs

26

Inc. v. Cyrix Corp., 42 USPQ2d 1872, 1875 (N.D. Cal. 1997) (emphasizing “importance of

27

precise definitions in the computer industry” in rejecting partial compatibility definition similar to

28

Google’s, and holding that compatible product “must support the same functions”).
ORACLE’S MAY 14, 2012 COPYRIGHT LIABILITY REPLY BRIEF
CASE NO. CV 10-03561 WHA
pa-1528434

9

1

Google claims its narrower definition “is not a position adopted just for this litigation.”

2

ECF No. 1116 at 17. That is false. Google acknowledged when it released the Android SDK in

3

November 2007 that Android is not compatible with Java. See TX 383 at 8 (Android Press Q&A)

4

(“048. Does Android support existing Java apps? / A. No. / 049. Is Android Java compatible? / A.

5

No.”). In contrast, Oracle’s definition of compatible has been used in the specification license

6

since Java’s release. See TX 980 at 6 (1996 API book); TX 610.1 (specification license).

7

The accepted definition of compatibility—and the relevant definition for this case—is that

8

programs written for Java will run on Android and vice versa. Android is incompatible. See ECF

9

No. 1118 at 18-19. But even under Google’s narrow definition, Android is not compatible.

10

Android does not have all of the classes and methods defined in these 37 packages in J2SE 5.0.

11

ECF No. 1124. The classes Google did include are nearly identical. For example, the package

12

java.security.auth.login in Android consists in its entirety of one exception, LoginException,

13

which has two constructors, LoginException() and LoginException(String message). This

14

exception, with the same two constructors, is present in the java.security.auth.login package in

15

Java as well. TX 610.2, TX 767. But any code that uses a class or method defined in one of the

16

37 packages in J2SE 5.0 but not in Android will not work on Android.

17

Questions 14-15: Inheritance Among Packages and Classes

18
19

The parties essentially agree on the rules of inheritance for packages and classes.
Question 16: SSO and the Java Language

20

The Court asked the parties to identify what Google copied other than names and input-

21

output designations. Google’s list is incomplete, so Oracle refers the Court to its list. See ECF

22

No. 1118 at 19-20. Google’s list does include exceptions (see ECF No. 1116 at 18), which Oracle

23

mentioned elsewhere in its brief. That should be added to Oracle’s response as well.

24

The Court also asked the parties, “to what extent was Android’s SSO dictated by the rules

25

of the basic programming language.” ECF No. 1088. Google ducked the question. See ECF

26

No. 1116 at 18. That is because the language does not dictate the SSO of either Android or Java.

27

Google’s expert conceded this. See RT at 2223:3-20 (“Classes have to be classes and packages

28

have to be packages, but what the functionality is, is what the API designers decide.”).
ORACLE’S MAY 14, 2012 COPYRIGHT LIABILITY REPLY BRIEF
CASE NO. CV 10-03561 WHA
pa-1528434

10

1

Dated: May 14, 2012

MORRISON & FOERSTER LLP

2
3

By: /s/ Michael A. Jacobs
Michael A. Jacobs

4

Attorneys for Plaintiff
ORACLE AMERICA, INC.

5
6
7
8
9
10
11
12
13
14
15
16
17
18
19
20
21
22
23
24
25
26
27
28
ORACLE’S MAY 14, 2012 COPYRIGHT LIABILITY REPLY BRIEF
CASE NO. CV 10-03561 WHA
pa-1528434

11