decoration decoration
Stories

GROKLAW
When you want to know more...
decoration
For layout only
Home
Archives
Site Map
Search
About Groklaw
Awards
Legal Research
Timelines
ApplevSamsung
ApplevSamsung p.2
ArchiveExplorer
Autozone
Bilski
Cases
Cast: Lawyers
Comes v. MS
Contracts/Documents
Courts
DRM
Gordon v MS
GPL
Grokdoc
HTML How To
IPI v RH
IV v. Google
Legal Docs
Lodsys
MS Litigations
MSvB&N
News Picks
Novell v. MS
Novell-MS Deal
ODF/OOXML
OOXML Appeals
OraclevGoogle
Patents
ProjectMonterey
Psystar
Quote Database
Red Hat v SCO
Salus Book
SCEA v Hotz
SCO Appeals
SCO Bankruptcy
SCO Financials
SCO Overview
SCO v IBM
SCO v Novell
SCO:Soup2Nuts
SCOsource
Sean Daly
Software Patents
Switch to Linux
Transcripts
Unix Books

Gear

Groklaw Gear

Click here to send an email to the editor of this weblog.


You won't find me on Facebook


Donate

Donate Paypal


No Legal Advice

The information on Groklaw is not intended to constitute legal advice. While Mark is a lawyer and he has asked other lawyers and law students to contribute articles, all of these articles are offered to help educate, not to provide specific legal advice. They are not your lawyers.

Here's Groklaw's comments policy.


What's New

STORIES
No new stories

COMMENTS last 48 hrs
No new comments


Sponsors

Hosting:
hosted by ibiblio

On servers donated to ibiblio by AMD.

Webmaster
Oracle v. Google - How to Proceed on the Copyright Issue III
Thursday, November 10 2011 @ 02:30 PM EST

We now have Oracle's response (611 [PDF] to the Google précis requested by the court, and the Oracle response appears to miss the mark at least as badly, if not more so, than Oracle claims Google missed the mark.

The Oracle errors are evident in the very first paragraph when Oracle states:

The Court asked Google to analyze concrete examples from at least two APIs to support Google’s proposal for an advance ruling on the 37 APIs through analytic dissection.


Although the court did ask for the two examples, those examples were only requested to support Google's answer to the following:
[T]he Court granted Google leave to file “a more developed précis . . . . that more fully develops its request for an advance determination regarding the 37 API specifications and the issue of their selection, arrangement, and structure.”
What the court appeared to be focused on with this request to Google is determining how the court should address whether Oracle had a protectible interest in the selection, arrangement and structure of the APIs. Implied in that request is that the court recognizes, on its face, that Oracle does not have a protectible interest in some, if not all, of the specific APIs, only in how they were selected, arranged and structured in Java.

In responding to the Google précis, what Oracle [Text] has principally done is rehash all of the same arguments it made in its earlier précis, which was rejected by the court. Oracle does not respond, point by point, to the Google arguments.

Let's start with the requested examples. In responding to the court Google stated:

Google discusses examples involving the selection of APIs for three classes, one method, and one field that are part of the java.net package. Google also discusses examples involving the hierarchy of the APIs for four classes that are part of the java.net package, and that are subclasses (or in one instance a subclass of a subclass) of the Object class in the java.lang package.
Google is providing examples addressing the issues of selection and arrangement and why it either did not copy Oracle or why it had no choice in order to properly interface with the Java language.

Did Oracle respond to these examples? No, it chooses its own examples and wants to argue them:

Two examples—java.util and java.nio—reflect significantly different design choices and illustrate the originality and creativity of the Java authors. The java.util package specification defines 49 classes, 16 interfaces, 1 enumeration, and 20 exceptions. Its diverse classes provide, among other things, calendar functions, random number functions, and collection manipulation functions—none of which have any necessary relationship to one another. (There is no technical or functional requirement that the calendar and collection classes be in the same package, for example.) Yet the API designers decided that these unrelated classes should be put in the same package.
Not only does Oracle not attempt to respond to or refute the examples provided by Google, it then provides an example that fails to prove the very point that Oracle wants to make, i.e., that there was something "expressive" about its selection. How can that be the case when, as Oracle states, "the API designers decided that these unrelated classes should be put in the same package." That makes it sound like things were randomly packaged together, not for a reason but merely for convenience or because they had to go somewhere. Is that expression? Keep in mind that Google was working to assure compatibility with the Java language, and Oracle has already conceded that Google had every right to use the Java language.

Oracle also asserts that the APIs were distinguishable and protectible because of the uniques way they were put together to be an attractive and efficient method for solving a problem for developers:

What distinguishes a great API from a good API is not its function, but its elegance and ease of use, its intuitive appeal, and its communication of substantive meaning and direction to developers. The authors of the Java APIs aimed for all three, and the success of the Java platform is a testament to their creativity.
Just because an item of software code embodies some creativity, that creativity does not trump the functional aspects of that code. This was established in both the Computer Associates case (Computer Assoc. Int'l v Altai, 982 F.2d 693, 714 (2d Cir. N.Y. 1992)) and Lotus case (Lotus Dev. Corp. v. Borland Int'l, 49 F.3d 807, 819 (1st Cir. 1995)).

In addition, the Oracle argument is inconsistent with the merger doctrine, which may apply even where there are multiple ways to implement a particular idea, but because of efficiency concerns, one or two choices become compelling such that they eliminate the use of any other form of expression. As stated in Nimmer on Copyright (4-13 Nimmer § 13.03[F][2]):

A great deal of computer science research has been devoted to developing methods of sorting or searching through data, and to analyzing the relative efficiency of various methods. As a result of such research, it is now recognized that some methods of sorting or searching are significantly more efficient than others in handling particular types of dta, even though any of numerous methods will work. In such cases, the merger doctrine should be applied to deny protection to those elements of a program dictated purely by efficiency concerns.
This is the point Google drives home in Section 3 of its précis.

Oracle's obsession with Google's compatibility argument is misplaced. Oracle wants the court to believe Google is talking about compatibility with either J2ME or JVM:

As a result of Google’s copying of only the parts of Java it wanted, Java programs that rely on the elements Google did not implement will not run on Android, and Android programs will not run on Java-compliant virtual machines. If Google had truly intended existing Java programs to be compatible with Android, it would have implemented the complete set of Java APIs and certified Android as Java-compatible using Oracle’s Technology Compatibility Kit.
However, Google is talking about compatibility with the Java language:
[I]n order for existing code in a language to be compatible and interoperable with new software written in the same language, the API elements that constitute the language must also be present, and named and organized identically.
We could go on, but the bottom line is that Oracle simply diverged on their own train of thought, repeating virtually everything they had argued in their earlier, rejected précis, while failing to respond to the specific arguments set forth by Google, which arguments were responsive to the court's request.

This same sort of rehashing of prior arguments is also evidenced in Oracle's reply to Google's opposition to exclude portions of the Leonard and Cox reports. (612 [PDF; Text]) Simply nothing new here.

In the final significant filing of the day Judge Alsup memorialized his rationale for appointing a Rule 706 expert (Dr. Kearl). ( 610 [PDF; Text])


***************

Docket:

610 – Filed and Effective: 11/09/2011
MEMORANDUM OPINION
Document Text: MEMORANDUM OPINION REGARDING RULE 706 EXPERT. Signed by Judge Alsup on November 9, 2011. (whalc1, COURT STAFF) (Filed on 11/9/2011) (Entered: 11/09/2011)

611 – Filed and Effective: 11/09/2011
RESPONSE
Document Text: RESPONSE to re 601 Letter in Opposition to Google's Further Precis re Copyrightability by Oracle America, Inc.. (Jacobs, Michael) (Filed on 11/9/2011) (Entered: 11/09/2011)

612 – Filed and Effective: 11/09/2011
REPLY
Document Text: REPLY (re 558 MOTION to Strike ORACLE AMERICA INC.S MOTION TO EXCLUDE PORTIONS OF THE EXPERT REPORTS OF GREGORY K. LEONARD AND ALAN J. COX - REDACTED ) ORACLE AMERICA, INC.S REPLY TO GOOGLE INC.S OPPOSITION TO MOTION TO EXCLUDE PORTIONS OF THE EXPERT REPORTS OF GREGORY K. LEONARD AND ALAN J. COX filed byOracle America, Inc.. (Attachments: # 1 Exhibit EXHIBIT C TO THE DECLARATION OF FRED NORTON IN SUPPORT OF ORACLE AMERICA, INC.S REPLY TO GOOGLE INC.S OPPOSITION TO MOTION TO EXCLUDE PORTIONS OF THE EXPERT REPORTS OF GREGORY K. LEONARD AND ALAN J. COX)(Holtzman, Steven) (Filed on 11/9/2011) (Entered: 11/09/2011)

613 – Filed and Effective: 11/09/2011
Sealed Document
Document Text: DOCUMENT E-FILED UNDER SEAL re 609 Order on Administrative Motion to File Under Seal EXHIBITS A & B TO THE DECLARATION OF FRED NORTON IN SUPPORT OF ORACLE AMERICA, INC.S REPLY TO GOOGLE INC.S OPPOSITION TO MOTION TO EXCLUDE PORTIONS OF THE EXPERT REPORTS OF GREGORY K. LEONARD AND ALAN J. COX by Oracle America, Inc.. (Holtzman, Steven) (Filed on 11/9/2011) (Entered: 11/09/2011)


*************

Documents:

610

IN THE UNITED STATES DISTRICT COURT
FOR THE NORTHERN DISTRICT OF CALIFORNIA

ORACLE AMERICA, INC.,
Plaintiff,
v.
GOOGLE INC.,
Defendant.

No. C 10-03561 WHA

MEMORANDUM
OPINION REGARDING
RULE 706 EXPERT

_______________________________

FRE 706(a) states that a district court “may appoint expert witnesses of its own selection.” The Supreme Court has long recognized the constitutionality of court-appointed experts. See, e.g., In re Peterson, 253 U.S 300, 312–13 (1920). The Court of Appeals for the Federal Circuit has acknowledged that “the Federal Rules do authorize those appointments,” and “under Ninth Circuit law, district courts enjoy wide latitude to make these appointments.” Monolithic Power Sys., Inc. v. O2 Micro Int’l Ltd., 558 F.3d 1341, 1346, 1348 (Fed. Cir. 2009).

In this patent and copyright infringement action, an independent damages expert was appointed pursuant to FRE 706 in accordance with the relevant case law, and this memorandum opinion explains why one was needed.

1. CIRCUMSTANCES WARRANTING APPOINTMENT OF RULE 706 EXPERT.

The appointment of a Rule 706 expert has been upheld “where the district court was confronted by what it viewed as an unusually complex case and what appeared to be starkly

conflicting expert testimony.” Monolithic, 588 F.3d at 1348. This action, too, is unusually complex and involves starkly conflicting expert views on the subject of damages.

The facts of this action have been set forth in previous orders (see Dkt. Nos. 137, 230, 433). Plaintiff Oracle America, Inc. originally asserted 132 claims from seven United States patents, as well as an expansive copyright-infringement claim (Dkt. No. 137 at 4). Oracle now intends to try 26 claims from six patents and several copyright-infringement theories to the jury (Dkt. No. 471 at 6). Oracle anticipates that this will take “a minimum of 28–30 hours” of evidence time (Dkt. No. 481 at 5). Defendant Google Inc. estimates that its defense to the liability portion of the case will require 150–175 trial hours for the patent claims alone, and “an additional non-trivial amount of time” for the copyright portion of the case (Doc. No. 480 at 7). Without acquiescing in those estimates, this order finds the obvious, namely that the disputed facts span several fields of computer science and involve complex technologies and products.

The damages aspect of this controversy is particularly involved. Damages from patent infringement are governed by different legal standards than damages from copyright infringement, but some accused products and acts are relevant to both parts of the action. The accused items are not entire products but rather elements of products, whose roles and relative importance within the larger units are disputed. The damages question is further complicated by the fact that the parties employ elaborate, nontraditional business models for distributing and monetizing the relevant products. For example, Google allegedly distributes its accused Android software free of charge, hoping to later benefit from improved market position and advertising revenue generated by Google searches on Android devices. Oracle, for its part, claims to have been harmed by the supposed fragmentation of its Java platform and developer community due to Google’s allegedly selective use of Java elements in Android.

Predictably, the parties and their retained economists have advanced extremely divergent views on the question of damages. The first damages study provided by Oracle’s economic expert opined that the fair market value of a hypothetical license for the alleged infringement would have been “at least $1.4 billion” and “could be as much as $6.1 billion” (Dkt. No. 230 at 4). In response, Google argued for zero or in the alternative stated that the total value of a

2

real-life licensing offer it rejected — “at most, a figure around $100 million” — should serve as a damages ceiling (id. at 10). The gap between these contentions was enormous. In light of the parties’ extremely divergent views on damages and the unusual complexity of the damages aspect of this case, an independent economic expert was needed to aid the jury.

2. PROCEDURES FOR SELECTION AND APPOINTMENT OF RULE 706 EXPERT.

The Court of Appeals for the Federal Circuit upheld the appointment of an independent expert by a district court where it found that “the district court properly administered the standards set by Rule 706.” Monolithic, 588 F.3d at 1347. FRE 706(a) provides the following procedural guidance:

The court may on its own motion or on the motion of any party enter an order to show cause why expert witnesses should not be appointed, and may request the parties to submit nominations. The court may appoint any expert witnesses agreed upon by the parties, and may appoint expert witnesses of its own selection. An expert witness shall not be appointed by the court unless the witness consents to act. A witness so appointed shall be informed of the witness’ duties by the court in writing, a copy of which shall be filed with the clerk, or at a conference in which the parties shall have opportunity to participate. A witness so appointed shall advise the parties of the witness’ findings, if any; the witness’ deposition may be taken by any party; and the witness may be called to testify by the court or any party. The witness shall be subject to cross-examination by each party, including a party calling the witness.
The Monolithic decision noted that the district court “allowed the parties to show cause why an expert witness should not be appointed” and then “instructed the parties to nominate candidates and confer upon a mutually agreeable witness.” It further noted that “the district court provided detailed written instructions to [the expert] regarding his duties, and ordered [the expert] to make himself available for depositions and for examination at trial.” In addition, it observed that appointment of the Rule 706 expert “did not limit in any way the parties’ ability to call their own experts,” and that the trial court allowed the parties’ experts to attack, support, or supplement the testimony of the court-appointed expert. Monolithic, 588 F.3d at 1347.

All of these procedures have been followed in the instant action (or will be inasmuch as we are still in process in the pretrial stage). In July 2011, the parties were apprised of the Court’s interest in appointing a Rule 706 damages expert “to testify before the jury at trial under FRE 706

3

and not [to serve] as a confidential advisor to the judge.” The parties were invited to show cause why such appointment would or would not be advisable and to describe the assignments that would be preferable. Counsel were requested to agree on two qualified candidates in this district, jointly call the candidates to make sure they would be available and have no conflicts, and then submit the candidates’ names and resumes to the Court (Dkt. No. 195).

Counsel initially were unable to agree on any joint candidate, so the geographical restriction was removed and the parties were invited to separately submit their own nominations (Dkt. No. 236). The parties ultimately found two mutually agreeable candidates for the position of independent damages expert. In August 2011, a conference was held with counsel and the two jointly nominated candidates for the purpose of vetting the candidates and discussing a proposed assignment-scope description which had been provided to counsel two weeks prior (Dkt. No. 273, 333). After the conference, the candidates were requested to provide follow-up information regarding possible conflicts and compensation (Dkt. No. 330).

After reviewing all relevant materials, Dr. James R. Kearl was appointed to serve as the Rule 706 damages expert. Attorney John Cooper was simultaneously appointed to serve as his pro bono counsel (Dkt. No. 374). The parties previously had been given notice that attorney Cooper agreed to represent the Rule 706 expert on a pro bono basis, and the parties were allowed to object to Attorney Cooper serving in this role (Dkt. No. 272). No objections were filed.

Attorney Cooper was instructed to work with counsel from both sides to draft a proposed order describing the scope of Dr. Kearl’s assignment and the procedures by which the parties would pay for his work and expenses (Dkt. No. 374). The first proposed order that was submitted required some revision (Dkt. No. 394). A revised version was signed into effect in September 2011 (Dkt. No. 413).

The assignment order contained a detailed description of Dr. Kearl’s assignment. It provided that Dr. Kearl would prepare an expert report and sit for deposition by stated deadlines. It also provided that each party would be able to examine Dr. Kearl at trial as though he were an adverse expert witness. The assignment order did not limit the ability of the parties to allow their experts critiqued by Dr. Kearl to respond at trial to his critiques. The order explicitly left open

4

the question of how Dr. Kearl’s role at trial would be described to the jury (Dkt. No. 413). Of note, the appeals court in Monolithic did “not perceive an abuse of discretion in the district court’s disclosure of [the expert’s] independent status to the jury” where it was accompanied by an instruction that the jury should not assign the court-appointed expert’s opinion greater inherent weight on account of his independent status. Monolithic, 588 F.3d at 1347.

To accommodate Dr. Kearl’s request for more work time, the damages issues were tentatively bifurcated from the liability trial, which at the time of Dr. Kearl’s appointment was scheduled to begin on October 31, 2011, a date that wound up being eclipsed by the need for the judge to preside over a long criminal trial. Dr. Kearl was instructed to be ready for a damages trial beginning in early December at the earliest (Dkt. No. 374). At a case management conference in October 2011, Dr. Kearl confirmed that he could complete his original assignment on the original time frame and also agreed to undertake a further analysis over the next three months. His deposition and report due dates were postponed accordingly (Dkt. No. 575). Although this is a particularly complex case, the scope of Dr. Kearl’s assignments have been appropriately matched to the amount of time available. See Monolithic, 588 F.3d at 1347 (finding that where a court-appointed expert was given only 35 days to review all relevant materials and prepare a report, the “strictures of this time frame” did not amount to an abuse of discretion).

3. COMPENSATION OF RULE 706 EXPERT.

The Federal Rules of Evidence provide that court-appointed expert witnesses “are entitled to reasonable compensation in whatever sum the court may allow.” In a civil action such as this, “the compensation shall be paid by the parties in such proportion and at such time as the court directs, and thereafter charged in like manner as other costs.” FRE 706(b). The Monolithic court upheld a district court’s appointment of an expert where “the court instructed the parties to share [the expert’s] reasonable fees and expenses.” Monolithic, 588 F.3d at 1347.

Here, the parties were warned in July 2011 that they would be required to pay the costs of any expert appointed pursuant to FRE 706 (Dkt. No. 236). The September 2011 order detailing Dr. Kearl’s assignment, which counsel for both sides helped Attorney Cooper draft, provided that each party would pay one half of Dr. Kearl’s monthly bills. The parties also reserved the right to

5

negotiate reductions in Dr. Kearl’s bills through Attorney Cooper, who was tasked with coordinating all payment matters (Dkt. No. 413 at 3). This arrangement was agreeable to counsel for all parties and the expert.

* * *

In sum, Dr. Kearl was selected and appointed in accordance with the Federal Rules of Evidence and the guidance of our courts of appeals. The assistance of an independent economic expert was and remains necessary to aid the jury in this complex technical action in light of the parties’ vastly divergent views on damages. The scope of Dr. Kearl’s assignment and the arrangements for his compensation were crafted so as not to interfere with the parties’ abilities to make their own presentations of the case.

Finally, the Court wishes to acknowledge again the exemplary role of Attorney John Cooper in volunteering his valuable services without compensation. John Cooper is an esteemed member of the bar and continues to serve with distinction in the highest tradition of this district.

Dated: November 9, 2011.

/s/ William Alsup
WILLIAM ALSUP
UNITED STATES DISTRICT JUDGE

6


611

[Morrison Foerster letterhead]

November 9, 2011

The Honorable William H. Alsup
United States District Court, Northern District of California
450 Golden Gate Ave., Courtroom 8, 19th Floor
San Francisco, CA 94102

Re: Oracle America, Inc. v. Google, Inc., Case No. 3:10-cv-03561-WHA

Dear Judge Alsup:

The Court asked Google to analyze concrete examples from at least two APIs to support Google’s proposal for an advance ruling on the 37 APIs through analytic dissection. Apparently realizing it isn’t seeking analytic dissection after all, Google nowhere performs this analysis.

Instead, with respect to every single aspect of the APIs it discusses, all Google says is that it had to implement those precise elements in order to achieve compatibility. Google already made this argument on summary judgment and lost. Google is not asking this Court to engage in an analytic dissection analysis of the type conducted in Apple v. Microsoft. It is asking to reargue its failed motion.

While Google frames its précis as a request for the Court to make an advance determination of copyrightability, it actually is asking for something different. Google’s professed desire for compatibility (which only lasts until Google says it doesn’t actually want to be fully compatible) goes more to its affirmative defense of fair use — a defense the Court has already held raises factual issues for trial—than to the copyrightability of the APIs themselves.

Google does not explain why its request will save the Court time. It will not. The Court has already held that the API design specifications are not unprotectable methods of operation. Oracle expects to prevail on copyrightability, but even if Google were to show that some or all elements of the API specifications are subject to merger or scenes a faire, Oracle would still present witnesses to show Google is liable for infringement because it made “virtually identical” copies.

The Honorable William H. Alsup
November 9, 2011
Page Two

I. THE JAVA API DESIGN SPECIFICATIONS

Copyright law protects expression in software design, including the selection and structure of software elements. The 37 Java API design specifications at issue represent years of creative effort in designing software that offer powerful features to developers. An API designer faces many creative choices: for example, which classes and interfaces to include (and whether a particular element should be a class or an interface), which fields and methods to include in each class or interface, the relationship of each element to other elements, and how to name and organize all of the elements in an elegant and easy-to-learn way. The choice of what to leave out—which classes should be left for programmers to implement, which fields and methods should be declared “private” and therefore part of a particular implementation but not the API— is creative as well. If the designer includes too little in an API, the developers will not have the tools and flexibility they prefer. If the designer includes too much, the APIs become overwhelming and nonintuitive.

Two examples—java.util and java.nio—reflect significantly different design choices and illustrate the originality and creativity of the Java authors. The java.util package specification defines 49 classes, 16 interfaces, 1 enumeration, and 20 exceptions. Its diverse classes provide, among other things, calendar functions, random number functions, and collection manipulation functions—none of which have any necessary relationship to one another. (There is no technical or functional requirement that the calendar and collection classes be in the same package, for example.) Yet the API designers decided that these unrelated classes should be put in the same package. See http://download.oracle.com/ javase/1.5.0/docs/api/java/util/ package-summary.html.

The API designers took a different approach to java.nio. The term “nio” stands for “New Input/Output,” and was born of a project led by Sun to create an improvement over the then existing Java I/O API, the java.io package. Although Sun decided that its new API should define classes implementing buffers, character sets, channels, and selectors, Sun chose to put only the

The Honorable William H. Alsup
November 9, 2011
Page Three

buffer-related classes into java.nio, and put the other classes into subpackages. Sun could have done it much differently: it could have augmented the java.io package to include all the new classes, or it could have put all the new classes into java.nio, rather than making separate subpackages. These choices were creative, not driven solely by function. In choosing to copy the selection and arrangement of the java.nio and java.util packages (as well as 35 other core Java API design specifications), nearly all in their entirety, Google infringed on a massive basis.

The API elements are arranged in a primarily hierarchical relationship, based on their inheritance from the interfaces they implement and the classes they extend. An example of class inheritance—and of Google’s copying of inheritance— is shown in these excerpts from the java.util specifications for Java and Android for the class “Hashtable”:

According to Oracle’s java.util API specification, the Hashtable class is a subclass of Dictionary, and implements three interfaces: Serializable, Cloneable, and Map, which are found in the java.lang, java.io, and java.util packages, respectively. This organization is hardly intuitive or preordained, but Google copied it. Android’s Hashtable class is likewise a subclass of Dictionary, and implements the same three interfaces, with the same three names, found in the same three corresponding Android packages.

The exceptions defined in the java.nio specification provide another example of the hierarchical choices made by the designers. Whereas BufferOverflowException was chosen to

The Honorable William H. Alsup
November 9, 2011
Page Four

be a subclass of java.lang.RuntimeException, InvalidMarkException was chosen to be a subclass of java.lang.IllegalStateException. In making these choices, the API designers expressed to developers that these two exceptions were different kinds of errors that should be categorized differently. The choice represents a creative expression with substantive meaning to developers, not something inalterably driven by function. And Google copied the organization and the names—not just in a substantially similar manner, but in a virtually identical form:

A third example is the Java Collections Framework, which is defined in the java.util API specification. The organization and hierarchy of the collections classes in Java is very different from that in other object-oriented languages, such as Smalltalk and C++, although they serve the same function. Google copied this unique organization as well. (See ECF No. 341, Ex. 1 ¶¶ 193-199 & Ex. 4.)

These are only three examples of expressive choices in the API specifications. Google repeatedly acknowledges it intentionally copied from Oracle’s copyrighted API specifications,

The Honorable William H. Alsup
November 9, 2011
Page Five

including the selection, arrangement, and structure of the API elements. (See Google Précis at 7- 15.) This copying includes thousands of names, classes, subclasses, interfaces, fields, methods, and exceptions. In the case of the java.util package alone, Google copied 49 classes, 16 interfaces, 20 exceptions, 85 fields, and 369 methods, and the structural relationships among them, as well as their structural relationships to the elements of other packages.

The selection, arrangement, and structure of API elements are an expression of the ideas underlying the API specifications. The creative effort that produced the specifications readily surpasses the “minimal” amount of originality required for copyrightability. See CDN Inc. v. Kapes, 197 F.3d 1256, 1260-61 (9th Cir. 1999). Google cannot seriously contest this. Both Google’s engineers and Google’s copyright expert have acknowledged that designing an API requires skill and creativity. As Google’s expert testified in deposition:

Q. Okay. And what are the elements of creativity involved in designing APIs?

A. I think that those are the same things that are the elements that – of creativity that go into designing. Other things related to software. And that’s kind of understanding how the software is meant to function.

And if you have a good understanding of how the software is going to function, and if you know the intended audience, that helps you to create things that that audience will like, will be able to use.

So the creativity is kind of understanding how the thing you’re going to design is going to be used. And understanding the audience or the libraries that you’re going to develop an API for, the creative aspects come into can you actually get at the needs or the wants or the usage of who your audience is.

(Astrachan Dep. 128:23-129:14.) Dr. Astrachan also testified:

Just as it’s hard to find people that are really good at anything that’s hard, whether it be, you know, being an artist, a football player, a concert violinist. Those things are hard. This [API design] is something that’s hard in the same way. (Id. at 128:9-13.)
What distinguishes a great API from a good API is not its function, but its elegance and ease of use, its intuitive appeal, and its communication of substantive meaning and direction to

The Honorable William H. Alsup
November 9, 2011
Page Six

developers. The authors of the Java APIs aimed for all three, and the success of the Java platform is a testament to their creativity.

II. GOOGLE IS ASKING THE COURT TO DECIDE ISSUES THAT IT SHOULD
BE REQUIRED TO PROVE AT TRIAL

Google’s précis does not propose that the Court engage in the type of analytic dissection of individual elements the district court performed in Apple v. Microsoft. Instead, despite being directed to provide specific examples, Google contends it should be given a blanket excuse for its copying because it wanted to make Android’s APIs compatible with the 37 Java APIs it copied. (See, e.g., Google Précis at 7-10, 12-15.) Google already made this sweeping argument in its summary judgment motion, and the Court rejected it. (See ECF No. 260, Google SJ motion at 14-17, 19-22; ECF No. 433, Summary Judgment Order (“SJ Order”) at 9, 12-13.) Now Google tries again, suggesting the parties should “update the briefing and declarations they submitted in connection with Google’s copyright summary judgment motion….” (Google Précis at 5.) Google is not entitled to a second bite at the summary judgment apple.

Moreover, while Google frames its précis as asking the Court to decide the copyrightability of the selection, arrangement and structure of the 37 API specifications (Google Précis at 1), in reality it is not. Google’s Ninth Circuit authorities, Sega and Sony, both addressed the defendants’ desire for compatibility in the context of evaluating fair use, not copyrightability. See Sega Enters. Ltd. v. Accolade, Inc., 977 F.2d 1510, 1520-28 (9th Cir. 1993); Sony Computer Entm’t, Inc. v. Connectix Corp., 203 F.3d 596, 602-09 (9th Cir. 2000).

Google should be required to try to prove its affirmative defense of fair use at trial if it can, and its desire for compatibility is, at best, only a part of one of at least four statutory factors that should be considered. (See SJ Order at 12 (quoting 17 U.S.C. § 107).) The Court has already found that Google’s fair use defense raises factual issues that preclude summary judgment in Google’s favor. (Id. at 12-13.)

The Honorable William H. Alsup
November 9, 2011
Page Seven

But even if Google had proposed that the Court perform a detailed analysis of individual elements of the API specifications in advance of trial, that would not be a useful exercise here. Even Google acknowledges analytic dissection is “helpful,” rather than mandatory. (Google Précis at 3 (quoting Apple Computer, Inc. v. Microsoft Corp., 35 F.3d 1435, 1443 (9th Cir. 1994).) Analytic dissection is not required here, and certainly not in advance of trial, where Google admits it engaged in wholesale, deliberate copying of literal and non-literal elements of the 37 API specifications, and where there are no issues of pre-existing license that pervaded Apple v. Microsoft. As the Tenth Circuit has noted, “Not every case requires an extensive abstraction-filtration-comparison analysis. Rather, ‘the appropriate test to be applied and the order in which its various components are to be applied . . . may vary depending upon the claims involved, the procedural posture of the suit, and the nature of the [works] at issue.’” Mitel, Inc. v. Iqtel, Inc., 124 F.3d 1366, 1372 (10th Cir. 1997) (citation omitted) (omissions in original); see also 4-13 Nimmer on Copyright § 13.03[A][1] (explaining need for extensive analytic dissection and similar tests arises only when there is “comprehensive nonliteral similarity”).

Oracle’s case is based on Google’s massive wholesale copying of literal and nonliteral elements, not copying of only nonliteral elements or a mere handful of literal elements. Substantial similarity will not turn on whether some of the elements individually could be filtered out. Google acknowledges that for nearly all of the 37 APIs it copied everything, so substantial similarity—and even virtual identity—is a given. See id. at § 13.03[A][2] (“Where there is literal similarity (virtually, though not necessarily, completely word for word) between plaintiff’s and defendant’s works … it is not necessary to determine the level of abstraction at which similarity ceases to consist of an ‘expression of ideas,’ because literal similarity by definition is always a similarity as to the expression of ideas.”).

Google is also incorrect in suggesting that courts decide copyrightability as a matter of law regardless of whether there are underlying factual issues for trial. See, e.g., Yue v. Chordiant

The Honorable William H. Alsup
November 9, 2011
Page Eight

Software, Inc., No. C 08-00019 JW, 2009 U.S. Dist. LEXIS 118824, at *7 (N. D. Cal. Dec. 21, 2009) (and cases cited therein) (finding triable issue as to whether plaintiff’s work was sufficiently original to warrant copyright protection). As Oracle stated in its trial brief, it believes that, after hearing the evidence presented at trial, the Court will decide copyrightability in its favor as a matter of law, but the Court may determine there are threshold questions of fact for the jury. (See ECF No. 536, Oracle Trial Brief at 23-24.)

Finally, Google’s argument that it is “crucial” for it to have further clarification on what Oracle is claiming was copied (Google Précis at 3) is meritless. Google copied the elements of the 37 API design specifications and the relationships among them, as illustrated in section I above, and reproduced them in its own specifications. It knowingly created a derivative work by basing its class library source code on the API design specifications, adopting Oracle’s selection, arrangement, and structure. Google acknowledges this copying. (See, e.g., id. at 7-15.)

III. GOOGLE’S COMPATIBILITY ARGUMENT IS LEGALLY INCORRECT, AND
PROVIDES NO BASIS FOR A DETERMINATION BEFORE TRIAL

Google’s proposal for an advance determination of its compatibility argument rests on a false factual premise and a misreading of the applicable decisions.

First, Android is not actually compatible with Java. Google fragmented Java by implementing less than the full number of the Java API packages (“subsetting”), adding its own APIs (“supersetting”) and, in some cases, only supporting portions of the classes within a Java API package. (See ECF No. 341-2, Mitchell Copyright Opp. Report ¶¶ 61-68, 101-113.) For example, Google chose to implement only 2 of the 16 classes in the java.awt.font package. As a result of Google’s copying of only the parts of Java it wanted, Java programs that rely on the elements Google did not implement will not run on Android, and Android programs will not run on Java-compliant virtual machines. If Google had truly intended existing Java programs to be

The Honorable William H. Alsup
November 9, 2011
Page Nine

compatible with Android, it would have implemented the complete set of Java APIs and certified Android as Java-compatible using Oracle’s Technology Compatibility Kit.

Google argues that “to ensure compatibility with the Oracle API packages that Google did implement in Android, Google had to use the same selection of API elements, and the same hierarchical arrangement and structure for those elements.” (Google Précis at 8 (emphasis in original).) The Court should reject this circular, selective compatibility argument—which even Google recognizes sounds “tautological.” (Id. at 14.) In essence, Google is saying that it wanted to copy certain features, so its infringement is justified because in order to achieve compatibility with these features, Google needed to copy them.

Google’s claim that it copied only what it had to in order to achieve compatibility is also false. (Id. at 7-8.) In many instances Google copied features that were not required for compatibility. For example, while Google may claim it needed to copy the types and orders of parameters in a method signature in order to achieve compatibility, it was not necessary for Google to use the same parameter names. (See ECF No. 391, Astrachan Rebuttal Report ¶ 42.) Where Google did copy Oracle’s parameter names, it did so because it wanted to. For example, the table below, excerpted from Prof. Mitchell’s Copyright Opening Report (ECF No. 341, Ex. 1 ¶ 205), shows method signatures and parameters for the class java.nio.IntBuffer. Google copied two of Oracle’s parameter names, “index” and “dst,” but modified two others, adding a prefix to “offset” and replacing “length” with “intCount.” Google could have used different names in place of “index” and “dst” as well, but instead Google copied Oracle’s name selection.

Google’s argument is also legally incorrect. Google misreads Sega and Sony. Both were “intermediate copying” cases that examined whether it was fair use to reverse engineer a

The Honorable William H. Alsup
November 9, 2011
Page Ten

copyrighted program where “disassembly is the only way to gain access to the ideas and functional elements embodied in a copyrighted computer program.” Sega Enters., 977 F.2d at 1527-28. The Ninth Circuit held that this intermediate copying “must have been ‘necessary’ to have been fair use.” Sony Computer, 203 F.3d at 603 (citing Sega, 977 F.2d at 1524-26). Sega and Sony did not hold, as Google suggests, that there is no infringement “if the implementing code was not copied.” (Google Précis at 6.) Rather, because the plaintiffs in those cases did not accuse the defendants’ finished products of infringement (unlike here), the court in Sega stated: “Our conclusion does not, of course, insulate [defendant] from a claim of copyright infringement with respect to its finished products.” Sega, 977 F.2d at 1528; see also Sony, 203 F.3d at 604 n.7 (final product not alleged to contain infringing material).

Unlike Sega and Sony, Oracle’s API specifications are posted online for anyone to see (though not to copy without a license), so there was no need to copy to discover their functional elements. Moreover, neither Sega nor Sony concerned rich, complex API specifications like the ones at issue here. These cases dealt with reverse engineering low-level functional requirements.

Sega and Sony’s holdings on fair use are distinguishable for another important reason. Copying in those cases was required to understanding functional ideas that were necessary for Accolade’s games to work on a Sega Genesis or for PlayStation games to work on Connectix’s emulator. That is not the case here, where Google selectively copied to attract programmers. Google did not have to copy Oracle’s designs. When asked whether Google could have written its own APIs for these 37 packages, Google’s expert Owen Astrachan replied: “It’s technically possible, but it’s not something that could have been done to realize the design goals that I mentioned earlier.” (Astrachan Dep. 253:2-4.) The design goals to which Dr. Astrachan referred were leveraging “a large body of software developers” and “a large existing body of code and other libraries.” (Id. 251:21-24.) This is fundamentally different from the functional compatibility that was required in Sega and Sony for the complementary products to perform.

The Honorable William H. Alsup
November 9, 2011
Page Eleven

Beyond the software arena, this Circuit has enforced copyrights even when the defendant faced strong pressures to adopt a copyrighted work as its own. In Practice Mgmt. Info. Corp. v. Am. Med. Ass’n, the Ninth Circuit upheld copyright protection for a coding system for medical procedures even where it became a federal agency’s standard, noting that there was nothing to stop the plaintiff’s competitors “from developing comparative or better coding systems and lobbying the federal government and private actors to adopt them.” 121 F.3d 516, 520 (9th Cir.1997); see also Del Madera Props. v. Rhodes & Gardner, Inc., 637 F. Supp. 262, 264 (N.D. Cal. 1985) (rejecting argument from property developers that governmental approval of copyrighted development plan justified use of map).

Limiting copyright protection for Oracle’s API specifications because they have become popular with programmers would punish Oracle for designing effectively. The Ninth Circuit reached a similar conclusion in examining this issue in the trademark context. See, e.g., Au-Tomotive Gold, Inc. v. Volkswagen of Am., Inc., 457 F.3d 1062, 1073 (9th Cir. 2006) (discussing analogous “aesthetic functionality” doctrine in trademark law: “We have squarely rejected the notion that ‘any feature of a product which contributes to the consumer appeal and saleability of the product is, as a matter of law, a functional element of that product.’”).

IV. THE MERGER DOCTRINE, SCENE A FAIRE, AND METHODS OF
OPERATION DO NOT SUPPORT GOOGLE’S PROPOSED ADVANCE
PROCEDURE.

Google points to the merger, scenes a faire, and methods of operation doctrines as supporting its proposed advance determination, but fails to make any showing that they apply. In its summary judgment order, the Court laid out Google’s requirements: “If Google believes, for example, that a particular method declaration is a scene a faire or is the only possible way to express a given function, then Google should provide evidence and argument supporting its views as to that method declaration.” (SJ Order at 9.) Google never does so here.

The Honorable William H. Alsup
November 9, 2011
Page Twelve

“Under the merger doctrine, courts will not protect a copyrighted work from infringement if the idea underlying the copyrighted work can be expressed in only one way, lest there be a monopoly on the underlying idea.” (Id. at 9 (quoting Satava v. Lowry, 323 F.3d 805, 812 n.5 (9th Cir. 2003)).) But Google does not give a single example of how there was only one, or very few, ways of expressing the APIs so that the merger doctrine might apply here. To the contrary, “Google does not dispute that, in at least many instances, Sun may have had other choices it could have made when it designed the APIs.” (Google Précis at 6.)

Google nonetheless contends it is not possible to distinguish the expression in the 37 API designs from the ideas of the APIs themselves. ( Id. at 9.) The claim is absurd. The API design specifications are comprised of thousands of elements. No copyright case has ever held this type of extensively detailed expression is simply an “idea.” The uncopyrightable ideas here may be the idea of a set of API packages or the idea of an API package containing a set of classes relating to input and output (e.g., java.io). But copyright protects Oracle’s particular selection, arrangement, and structure of the 50 classes, 12 interfaces, and 16 exceptions (and all of their constituent member fields and methods) that make up the java.io package, based on choices intended to convey meaning and content to developers. Google copied Oracle’s specific expression, not merely the underlying idea.

Google has similarly failed to make any showing that it merely copied scenes a faire. Google has again not provided any examples of API elements that consist of “commonplace expressions” that “are indispensable and naturally associated with the treatment of a given idea.” (SJ Order at 8-9.) Instead, Google argues that the scenes a faire doctrine covers everything Google copied because Google purportedly did so in the name of compatibility.

Google has scenes a faire law backwards. The proper test under scenes a faire is whether the copyrightholder, Oracle, was constrained in its choice of design, not Google. Control Data Sys., Inc. v. Infoware, Inc., 903 F. Supp. 1316, 1323 (D. Minn. 1995). Google miscites

The Honorable William H. Alsup
November 9, 2011
Page Thirteen

Computer Associates, which focused on the plaintiff’s copyrighted work. 982 F.2d at 710 (stating that court must “examine the structural content of an allegedly infringed program for elements that might have been dictated by external factors”) (emphasis added); see also Mitel, 124 F.3d at 1375 (lower court’s proper focus “should have remained upon the external factors that dictated [plaintiff’s] selection of registers, descriptions, and values,” rather than on whether external factors “justified [defendant’s] copying.”).

Google cites no contrary Ninth Circuit authority. The footnote Google cites from Sega focuses on the plaintiff’s primitive code, stating: “Sega’s key appears to be functional. It consists merely of 20 bytes of initialization code plus the letters S-E-G-A.” 977 F.2d at 1524 n.7. Nor do the two other cases cited in Goldstein on Copyright § 2.3.2.1 n.46 support Google. Atari Games Corp. v. Nintendo of Am. Inc., 975 F.2d 832 (Fed. Cir. 1992) actually holds the opposite. In Atari, the Federal Circuit applied Ninth Circuit law to uphold the plaintiff’s preliminary injunction based on Atari’s infringement of its copyright in the key used to unlock Nintendo’s video game console. Id. at 836-37, 847. Reasoning that the key contained a “unique sequence” of code, and that “External factors did not dictate the design” of Nintendo’s key, the Court found that the code was entitled to copyright protection. Id. at 840. Thus, while Atari could reverse engineer the code to learn its unprotected features - just like the defendants in Sony and Sega - it could not replicate the copyrighted code. Id. at 844. Goldstein also relies upon a quote from Apple Computer, Inc. v. Microsoft Corp., which merely recites different policy considerations in copyright law. See 799 F. Supp. 1006, 1025-26 (N.D. Cal. 1992) (referring to the “ambitious enterprise” of “[s]triking [a] balance” between copyright’s purpose of excluding free riders “who do not pay the cost of creation” with “overly inclusive copyright protection” that “can produce its own negative effects by inhibiting the adoption of compatible standards”).

The Honorable William H. Alsup
November 9, 2011
Page Fourteen

Google’s approach to methods of operation is similar. Rather than give a concrete example of a method of operation, it attempts to re-litigate its failed summary judgment argument, contending the APIs as a whole are a method of operation. (See Google Précis at 5-6.)

The parties have already argued this issue at length. Google’s earlier, fully briefed argument on methods of operation relied on Lotus Dev. Corp. v. Borland Int’l, Inc., 49 F.3d 807 (1st Cir. 1995), a decision that has never been adopted by the Ninth Circuit and that has expressly been rejected by others. See Mitel, 124 F.3d at 1372. Lotus’s holding that a “method of operation” is any “means by which a person operates something,” Lotus, 49 F.3d at 815, would strip copyright protection from many computer programs. Instead, the Ninth Circuit asks “whether, on the particular facts of each case, the component in question qualifies as the expression of an idea, or an idea itself,” Johnson Controls, Inc. v. Phoenix Control Sys., Inc., 886 F.2d 1173, 1175 (9th Cir. 1989). The Java API specifications are expression, not merely ideas.

V. ADOPTING GOOGLE’S APPROACH WILL BE TIME CONSUMING AND
WILL REQUIRE DUPLICATION AT TRIAL

Google devotes only two sentences to the question of whether its proposal will simplify issues for trial. It will not. Oracle believes there is no real question that the selection and arrangement of the Java API specifications exceed the minimal copyrightability requirements of Feist. If it prevails in this advance proceeding, Oracle will still need to explain the selection and arrangement of the API specifications to the jury to show the extent of Google’s copying to prove substantial similarity and copying and to support Oracle’s claim for damages. And Google, for its part, has not abandoned its fair use defense. Thus, if Google should lose on its uncopyrightable subject matter argument, the Court will be faced at trial with a reprise of the very same set of issues, albeit under a different legal rubric.

The Honorable William H. Alsup
November 9, 2011
Page Fifteen

Conversely, even if Google were to prevail on all, or a significant portion of, its merger or scenes a faire arguments, the copyright claim would still go forward. Google would still be liable for copyright infringement if Oracle could show “virtual identity.” See Apple, 35 F.3d at 1439. Google proclaims that it “had to use the same selection of API elements, and the same hierarchical arrangement and structure,” purportedly to achieve compatibility (Google Précis at 8), and Oracle would be entitled to present this same evidence at trial.

Finally, Google’s argument rests on its purported need for compatibility. As discussed in section III above, there are serious factual problems with Google’s compatibility defense. The defense should be tested with live testimony and cross-examination, not decided on a cold record. Regardless of the outcome, if this proceeding does go forward, witnesses from both sides will then need to be presented to the Court a second time at trial. Google’s proposal will squander, not conserve, the Court’s time.

Respectfully submitted,

/s/ Michael A. Jacobs
Michael A. Jacobs


612

UNITED STATES DISTRICT COURT
NORTHERN DISTRICT OF CALIFORNIA
SAN FRANCISCO DIVISION

ORACLE AMERICA, INC.
Plaintiff,
v.
GOOGLE, INC.
Defendant.

Case No. CV 10-03561 WHA

ORACLE AMERICA, INC.’S REPLY TO
GOOGLE INC.’S OPPOSITION TO MOTION
TO EXCLUDE PORTIONS OF THE EXPERT
REPORTS OF GREGORY K. LEONARD AND
ALAN J. COX

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

Google defends Drs. Leonard and Cox largely by relying on two false premises: first, that it is an acceptable expert practice to adopt, without qualification or analysis, the post-discovery statements of Google employees (Opp. at 3, 4, 5, 7); and second, that the ends-based advocacy by Google’s experts is justified because Oracle’s experts supposedly offer “made-for-litigation” or “results-oriented” opinions. (Opp. at 1, 5–6, 12–13, 14–15.) Both arguments are wrong. Unlike Oracle’s experts, Drs. Leonard and Cox rely on interview statements that are contradicted by the record evidence, which they justify ignoring because contemporaneous documents must yield to what they now know, based on other witnesses for whom they vouch, law they misapply, and expertise beyond any they might legitimately claim. This approach is fundamentally subjective and scientifically unsound.

I. Drs. Leonard And Cox Rely On Spoon-Fed Facts From Google Employees In Ways That
Further Discovery Cannot Cure

The extraordinary extent of Drs. Leonard’s and Cox’s reliance on interviews is even clearer now that they have been deposed. Neither questioned the witnesses about what they had written in the past.1 (Declaration of Fred Norton (“Norton Decl.”) Ex. A (“Leonard Dep.”) at 97:13–23; Norton Decl. Ex. B (“Cox Dep.”) at 160:21–24.) Dr. Leonard’s role as scrivener was so ministerial that he simply typed everything he was told by Google employee Brian Swetland straight into his report. (Leonard Dep. at 93:10–24; 310:22–311:1.) When asked what he had done as an expert to evaluate the viability of alternatives, Dr. Leonard’s answer was that he knew how to ask “the right questions.” (Id. at 259:9– 21.) Whatever those questions were, once Google’s employees told him that switching would be costless, he considered his job “done” and “there’s no reason to go further.” (Id. at 259:9–21 (’720), 254:12–257:10 (’205/JIT).)

This abdication of the role of the expert cannot be cured by cross-examination. “[W]hile Rule 703 was intended to liberalize the rules relating to expert testimony, it was not intended to abolish the hearsay rule and to allow a witness, under the guise of giving expert testimony, to in effect become the mouthpiece of the witnesses on whose statements or opinions the expert purports to base his opinion.”

______________________________

1 Each witness was interviewed once, and with only one exception, for 45 minutes or less. (See Leonard Dep. at 310:18–314:19.) Drs. Leonard and Cox took notes for only a few interviews, and then usually only a cryptic page or two. (See id. (testimony about 8 handwritten pages of notes for all interviews); Cox Dep. at 160:21–167:25 (testimony about 15 handwritten pages of notes for all).)

1

Loeffel Steel Prods., Inc. v. Delta Brands, Inc., 387 F. Supp. 2d 794, 808 (N.D. Ill. 2005).

Drs. Leonard’s and Cox’s interview-based opinions repeatedly contradict documents written by the interviewees. Compare Cox Dep. at 181:4–181:15 with Norton Decl. Ex. C (GOOGLE-01- 00025575 at 584) (2006 wireless industry adoption of Java); Cox. Dep. at 275:17–278:1 and Leonard Dep. at 308:23–309:22 with Dearborn 10/21 Decl. Ex. 9 (Dkt. No. 585-1 at 20) (lack of Java fragmentation in 2006); Cox Dep. at 239:10–255:3 with Dearborn 10/21 Decl. Ex. 5 (Dkt. No. 585-1 at 5–7) (advantages of Java over C++); Leonard Dep. at 274:22–275:12 with Dearborn 10/27 Decl. Exs. B–E (Dkt. No. 573-1) (importance of speed). Reliance on interviews inconsistent with the evidence warrants exclusion of the testimony. See IP Innovation L.L.C. v. Red Hat, Inc., 705 F. Supp. 2d 687, 690 (E.D. Tex. 2010) (excluding opinion founded on “blatant oversight” and assumptions contrary to record evidence).

Google defends its experts’ cherry-picking by claiming that it “will, and accepts that it must, offer the underlying factual testimony from the percipient witnesses first, before its experts may testify based on those facts” (Opp. at 3); that “the interviews constitute facts and data ‘of a type reasonably relied by experts in the particular field,’” (id. at 4 (citing only Rule 703)); and that this Court has not held that “party employee interviews are an inherently unreliable basis for an expert opinion.” (Id.) None of these arguments justifies what Drs. Cox and Leonard did.

Google has not eliminated the hearsay problem. Its experts will offer opinions at trial based on information they were given during post-discovery interviews. When an expert ignores extensive record evidence to rely uncritically on interviews with “highly partisan” employees of the client, the opinion fails the standards for admissibility. Therasense Inc. v. Becton, Dickinson & Co., 2008 WL 2323856, at *2 (N.D. Cal. May 22, 2008) (“any opinion based on such untested and partisan foundation is not based on sufficient facts and data within the meaning of Rule 702”).2 Even if further additional

____________________________

2 The other two cases cited by Google for its reliance on interviews (Opp. at 4) do not rescue its experts either. Both cases involved accounting experts, and AICPA standards specifically allow accountants to base their opinions on interviews with their clients. See, e.g., Inline Connection Corp. v. AOL Time Warner, Inc., 470 F. Supp. 2d 435, 443 (D. Del. 2007) (quoting AICPA standards). Neither case holds that an economist may sit back and claim that party employees told them alternatives would be costless, as Google’s experts have done here. Additionally, Inline specifically noted that the experts’ opinion in that case (unlike the situation here) did not clearly contradict other factual evidence. Id.

2

discovery could fully cure this defect, Google refuses any new depositions for five of the seven interviewees. (See Opp. at 6–7; Norton Decl. Ex. D.) Google’s references to the depositions that have been taken (Opp. at 6–7) do not help, as they could not have addressed the accuracy of the experts’ paraphrases, their omission of bad facts disclosed by employees, or other aspects of the interviews.

II. Drs. Leonard And Cox Offer No Competent Economic Opinions On Alternatives

Drs. Leonard and Cox devote lengthy portions of their reports to technical opinions on noninfringing alternatives,3 despite their lack of expertise on that subject. (Oracle Br. at 4–5.) Google contends that Drs. Leonard and Cox actually provide only economic opinions based on those “foundational” facts. (Opp. at 1, 2.) Google is wrong.

Drs. Cox and Dr. Leonard provide no economic analysis of Google’s purported alternatives. Dr. Cox admitted at deposition that he did not account for costs that Google avoided as a result of infringement because he had “been told” that “the evidence will show, that there are not substantial costs, and possibly zero additional costs, that are avoided” by using a C-based language rather than Java. (Cox Dep. at 83:1–85:10.) The entire “economic analysis” supporting that conclusion appears in one sentence in Dr. Cox’s report: “The incremental cost and time to Google of going with C++ or another programming language other than the Java programming language would not have been significant.” (Cox Report at 27.) In support of that “analysis,” Dr. Cox cites only his interviews with three Google employees. (Id. at 28 n. 105.) Despite Google’s uncontroverted documents and testimony admitting that the launch window for Android was critical, Dr. Cox has no analysis of the economic impact of delay as a result of the supposed alternative. (Cox Dep. at 88:21–89:22.)

Similarly, Dr. Leonard does not quantify—by any metric—costs, risks, benefits, or market effects of Google’s purported alternatives. For example, for the “alternative” of using a native compiler instead of a virtual machine, he simply accepts the representations of Google employees that the out-ofpocket cost and time would have been the same for either choice. (Leonard Dep. at 212:11–24.) He does the same with regard to the “alternative” of using C++, accepting without analysis interviewees’

_________________________________

3 Dr. Leonard discusses Google’s purported non-infringing alternatives in his report on pages 10–30, 31, 37, 45, 48, 63, 69 (Georgia-Pacific factor 7), 70 (Georgia-Pacific factor 9), 71 (Georgia-Pacific factor 13), 72, 74, 75, 80, 88, 89, 94, 95, 97, 98, and 118. Dr. Cox discusses Google’s purported noninfringing alternatives in his report on pages 7, 15, 23–27, 38, 48, 59, 60, and 61.

3

statements that the added development time “would not have been significant.”4 (Id. at 36:13–39:24.) His patent-by-patent analysis was similarly superficial. (See, e.g., id. at 238:7–242:12 (no quantification of effect of removing ’702).)

Drs. Leonard and Cox do not merely “rely for technical background on the opinions of Google’s technical experts” (Opp. at 1); they try to bolster Google’s technical experts by citing interviews that they claim—but cannot competently testify—support those experts’ opinions and vouch for the technical witnesses that must be believed for their “economic opinions” to have a basis. That is impermissible. See, e.g., Dkt. No. 56 ¶ 12 (“experts lacking percipient knowledge should avoid vouching for the credibility of witnesses, i.e., whose version of the facts in dispute is correct.”); Loeffel, 387 F. Supp. 2d at 808 (“[a] scientist, however well credentialed he may be, is not permitted to be the mouthpiece of a scientist in a different specialty. That would not be responsible science.”).

III. Drs. Leonard and Cox Offer No Competent Economic Opinions As To “Fragmentation”

Drs. Leonard and Cox opine that Sun had little to lose by abandoning Java to Google, because Java was already “stagnant” or “fragmented.” Drs. Cox and Leonard have no expertise in assessing the vitality of a complex software platform, and they have applied no economic analysis to assess the effects of any stagnation or fragmentation of Java. Google admits as much. It defends Drs. Leonard and Cox by arguing that they rely on witnesses, documents, and testimony that supposedly say the same thing. (Opp. at 10–11.) But an expert “may not simply parrot or recite the opinions and knowledge of other expert and fact witnesses.” Ash Grove Cement Co. v. Emp’rs Ins. Co., 246 F.R.D. 656 at 661 (D. Kan. 2007); see also Stein v. Pacific Bell, 2007 WL 831750, at *11 (N.D. Cal. Mar. 19, 2007).

IV. Drs. Leonard And Cox Offer Opinions That Are Contrary To Law

Dr. Cox improperly reduces wrongful profits damages by arguing that Google could have

__________________________________

4 The experts also improperly apply ex post facts to ex ante hypothetical negotiations. For example, Google claims that Dr. Cox can claim that using a non-Java language was a viable alternative because of “massive success of Apple’s iPhone,” (Opp. at 10), but it is legally improper to use the ultimate success of the iPhone—which contains its own hardware and was launched in 2007—to ratify alternatives the parties might have considered in the hypothetical negotiation for a software platform in 2006. See, e.g., Lucent Techs., Inc. v. Gateway, Inc., 580 F. 3d 1301, 1333–34 (Fed. Cir. 2009) (describing proper use of later-dated evidence). Similarly, Dr. Leonard’s entire basis for concluding that Sun’s business model projections in 2006 are unreliable is hearsay within hearsay—a newspaper article reporting analyst report statements—from late 2008. (Leonard Dep. at 191:6-193:22).

4

earned those profits without infringing. Google defends that error by citing the copyright statute’s causation requirement. (Opp. at 13–14.) But Google cites no case holding that profits actually caused by the infringement may be ignored because the infringer could have chosen to sell a non-infringing product—but didn’t. The Copyright Act awards the profits that are “attributable to the infringement.” See Polar Bear Prods., Inc. v. Timex Corp., 384 F.3d 700, 707–08 (9th Cir. 2004) (causal-link standard; quoting 17 U.S.C. § 504(b)). “If the infringer has earned a profit, this award makes him disgorge the profit to insure that he not benefit from his wrongdoing.” On Davis v. The Gap, Inc., 246 F.3d 152, 159 (2d Cir. 2001). The causation test for this disgorgement remedy is not, as Dr. Cox assumes, whether the infringer had alternatives; such a requirement would gut the purpose of wrongful profits. (Oracle Br. at 13.) The Court should strike Dr. Cox’s wrongful-profits analysis.

Google’s mitigation argument is also wrong. Dr. Leonard claims the upward adjustment for Project Armstrong is “inappropriate because it is counting as ‘damages’ something that Sun could have mitigated,” so the adjustment should be “zero.” (Leonard Report at 83.) Google concedes it has no mitigation defense, but says Dr. Leonard is really testifying about causation. (Opp. at 13.) Dr. Leonard has no expertise that would allow him to conclude it was feasible to monetize the infringing Android. In any event, Oracle’s decision not to pursue a business with a Java-incompatible Android is irrelevant to its ability to monetize a Java-compatible Android. This testimony could only confuse the jury.

V. Dr. Leonard Lacks Expertise Necessary To Testify On Conjoint Analysis

Dr. Leonard purports to critique Dr. Shugan’s conjoint survey and analysis of consumer preferences for the attributes provided by the copyrights and patents in suit. As explained in Dr. Shugan’s declaration, submitted with Oracle’s opening brief, Dr. Leonard lacks the expertise necessary to critique Dr. Shugan’s work and fundamentally misunderstands the marketing literature on conjoint analysis. Dr. Leonard’s response is to file a declaration that simply misreads a larger number of articles, compounding his errors and confirming his lack of expertise. (Shugan Response Decl. ¶¶ 4– 22.) Dr. Leonard’s testimony on conjoint analysis should be excluded.

Dated: November 1, 2011

BOIES, SCHILLER & FLEXNER LLP

By: /s/ Steven C. Holtzman
Steven C. Holtzman

Attorneys for Plaintiff
ORACLE AMERICA, INC.

5



  


Oracle v. Google - How to Proceed on the Copyright Issue III | 123 comments | Create New Account
Comments belong to whoever posts them. Please notify us of inappropriate comments.
Off Topic here
Authored by: DannyB on Thursday, November 10 2011 @ 02:39 PM EST
Please post off topic comments here. Provide clickable links if possible.

---
The price of freedom is eternal litigation.

[ Reply to This | # ]

corrections thread here
Authored by: designerfx on Thursday, November 10 2011 @ 03:25 PM EST
corrections and spelling errors, etc.

[ Reply to This | # ]

newspicks thread
Authored by: designerfx on Thursday, November 10 2011 @ 03:26 PM EST
with link to newspick if it's the first comment

[ Reply to This | # ]

comes thread
Authored by: designerfx on Thursday, November 10 2011 @ 03:27 PM EST
comes documents here with html formatting, posted in plain old
text please

[ Reply to This | # ]

Oracle v. Google - How to Proceed on the Copyright Issue III
Authored by: cjk fossman on Thursday, November 10 2011 @ 03:50 PM EST
In responding to the Google précis, what Oracle has principally done is rehash all of the same arguments it made in its earlier précis, which was rejected by the court. Oracle does not respond, point by point, to the Google arguments.

Standard BS&F tactic.

Seems like Oracle should be renamed SCOracle or SCOredux.

[ Reply to This | # ]

Oracle v. Google - How to Proceed on the Copyright Issue III
Authored by: Anonymous on Thursday, November 10 2011 @ 04:07 PM EST
Fuzzy memory here, but regarding the following claim:

"If Google had truly intended existing Java programs to be compatible with
Android, it would have implemented the complete set of Java APIs and certified
Android as Java-compatible using Oracle’s Technology Compatibility Kit."

Wasnt one of the big reasons fror Android being an independent implementation
the fact that they wanted to license J2SE for mobile use but were only offered a
J2ME license? And there's a nagging memory that the real sticking point here was
that it was the TCK that Sun refused to license to the Android guys.

I'm probably completely wrong here... but hoping for enlightenment!

[ Reply to This | # ]

"No technical or functional requirement"
Authored by: bugstomper on Thursday, November 10 2011 @ 04:41 PM EST
Oracle says "There is no technical or functional requirement that the
calendar and collection classes be in the same package, for example".

While there may be no technical or functional requirement for the API designers
to have put them in the same package, once they have made that choice there is
no way to write a calendar class and a collection class that are interoperable
with programs written to that API without using the same names for the classes
and putting them in the same package.

The API designers created the technical and functional requirement when they
made their arbitrary choice.

As Mark pointed out, this brings the merger doctrine into play - Even if the
choice of which classes go into which packages, how they are named, etc., is a
matter of creative expression, once those elements of the API are specified
there is no other functional way to express those particular elements. That is
the ultimate example of there being one standard way to express the idea, which
is what the merger doctrine is all about.

[ Reply to This | # ]

Ridiculous
Authored by: Anonymous on Thursday, November 10 2011 @ 07:08 PM EST
According to Oracle’s java.util API specification, the Hashtable class is a subclass of Dictionary, and implements three interfaces: Serializable, Cloneable, and Map, which are found in the java.lang, java.io, and java.util packages, respectively. This organization is hardly intuitive or preordained, but Google copied it. Android’s Hashtable class is likewise a subclass of Dictionary, and implements the same three interfaces, with the same three names, found in the same three corresponding Android packages.
Those are called interfaces for a reason: they advertise that something offers certain features, making it compatible. That is to say, if they did not copy this, they would cause the very fragmentation that Oracle complains of.

Similarly, the class hierarchy is little different than saying that a human is a type of mammal, which in turn is a type of animal. Everything in Java derives from Object[1]. So that leaves you saying merely that a hash is a type of dictionary. That is akin to saying that humans are mammals. While one might suppose that other organizations are possible, there simply aren't many that make any sense.

In short, this example is one which undercuts their case. If this is the best they have to offer, Oracle deserves to be laughed out of the courtroom.

[1] There are also primitive types like ints which exist for performance reasons (as well as corresponding classes, Integer in this example). But the object shown here is not a primitive and they have no reasonable choice but to derive it from Object. For more information on boxing, see this doc hosted by Oracle.

[ Reply to This | # ]

60 Hz... meh!
Authored by: Anonymous on Thursday, November 10 2011 @ 09:34 PM EST
I prefer 400 Hz; you can get significantly more power out of
the same mass transformer. Not that it matters with
switchers, which operate at a much higher frequency, anyway.
I've built DC step-up systems for tube amplifiers, to take DC
from 12V to 250-300V, and the leakage at 400 Hz really isn't
an issue, if all is bypassed decently. And, you can get a
lot more kick out of the same weight iron core!

Regards,
mc

[ Reply to This | # ]

Wait - is Oracle complaining that Google copied Java badly?
Authored by: Anonymous on Friday, November 11 2011 @ 12:10 AM EST
"Yer honor, there's no way Google could have made Android compatible with
Java without copying Java. And look, it's invompatible with Java!"

(I know it's not the whole of the case and even technically accurate, but it's
odd seeing Oracle complaining about copying on the one hand, and incompatibility
on the other)

bkd

[ Reply to This | # ]

The world has definitely changed!
Authored by: Anonymous on Friday, November 11 2011 @ 12:16 AM EST
Ah, sorry, it was not on the main page. It was one of their (The Economist's)
blog postings, and it contained a media clip - either audio or video. It was
the media clip that gave me that message.

MSS2

[ Reply to This | # ]

Are Oracle being honest when they claim not to understand creative expression?
Authored by: Ian Al on Friday, November 11 2011 @ 05:19 AM EST
They would be incompetant to practice in the area of copyright law if they did not know that ideas are not protectable by copyright. Similarly, facts cannot be protected by copyright otherwise only one news agency would be able to report a particular news item. Copyright is to protect the creative expression in an expressive work like a book or a painting from being copied without a licence to do so.

[T]he Court granted Google leave to file “a more developed précis . . . . that more fully develops its request for an advance determination regarding the 37 API specifications and the issue of their selection, arrangement, and structure.
To paraphrase that for Oracle's benefit (although, they have always ignored my helpful analysis, in the past) the judge wanted Google to develop their contention that selection, arrangement and structure in the 37 APIs did not contain protectable creative expression.

Oracle repeated their original assertion about what made a great API.
What distinguishes a great API from a good API is not its function, but its elegance and ease of use, its intuitive appeal, and its communication of substantive meaning and direction to developers. The authors of the Java APIs aimed for all three, and the success of the Java platform is a testament to their creativity.
I suggest that the 'elegance and ease of use' of an API relates to the access of the user to the ideas and facts within the API and is not creative expression in its own right. Intuitive appeal must mean that the user intuitively accesses the ideas expressed by the API. Communication of substantive meaning and direction can only mean the communication of ideas to the user. The overarching theme is that all of the features referred to as 'great', by Oracle, are to do with the access offered to the user of the ideas, facts and functionality of the API. None of this is protectable creative expression.

Oracle fail to demonstrate that literary expression within the text of the APIs (the only vehicle in the APIs for creative expression) are the means by which the objective of being a great API are met.

That is completely excusable, since that is not what the judge was asking for. The judge was asking for a development of the contention that there was no protectable creative expression in the APIs by virtue of their selection, arrangement, and structure.

Oracle are quite justified in maintaining that selection, arrangement, and structure are time consuming parts of the creative endeavour of devising a great API that has the features of elegance and ease of use, intuitive appeal, and communication of substantive meaning and direction to developers.

However, their examples of a java package which contained a ragbag of functions do not demonstrate 'originality and creativity' in API design, but the complete opposite. Claiming that Google copied their lack of originality and creativity begs the question of whether Google had any choice. Google answer that they had none, if users were to be able to use the Java language.

Oracle know that software interface specifications have been the subject of similar discussion in past court cases. I think they have specifically referred to the point that APIs have not been universally ruled as not containing protectable creative expression. In each case, the APIs in question must be examined for protectable creative expression. That leads me to deduce that Oracle are being wilfully and serially dishonest in their refusal to acknowledge copyright law.

---
Regards
Ian Al
Software Patent: code for Profit!

[ Reply to This | # ]

Groklaw © Copyright 2003-2013 Pamela Jones.
All trademarks and copyrights on this page are owned by their respective owners.
Comments are owned by the individual posters.

PJ's articles are licensed under a Creative Commons License. ( Details )