Friday, October 9, 2020

FACT CHECK: Ten falsehoods and fallacies Google's lawyer told the Supreme Court about Oracle's Android-Java copyright case

I live-tweeted about the Google v. Oracle America oral argument before the Supreme Court, and about five minutes into the hearing I already felt that Google was likely to lose the copyrightability part. After all justices had indicated where they stood on that question, or what they were interested in, there was no more doubt to me that Oracle will win that part by a unanimous or near-unanimous decision--but "fair use" is harder to predict, with a remand to the Federal Circuit being a possibility. Yesterday I published the justices' copyrightability statements/questions and commented on them.

Timothy B. Lee, who opposes API copyrightability, wrote on Ars Technica that Google's Supreme Court faceoff with Oracle was a disaster for Google, and it appears the finger-pointing is already in full swing. Lots of law professors supported Google for ideological reasons and maybe because they were misled about the expressive and original nature of API declaring code. One of them, Cornell's James Grimmelmann, is quoted by Ars Technica as blaming Google's lead Supreme Court counsel, Thomas Goldstein, for having done "an abysmal job." I actually saw Mr. Goldstein at his best when he represented Qualcomm before a Ninth Circuit panel earlier this year, and I wouldn't attribute to his performance on Wednesday an outcome that can and actually must be explained with the spuriousness of Google's non-copyrightability argument. Of course, one could have tried to take a different angle on the issue, and Professor Grimmelmann would have preferred Mr. Goldstein to make a more coding-centric argument. Considering how the justices approached the subject, however, Mr. Goldstein's strategy might have been the better choice--just that declaring code is program code, and program code is copyrightable if it's original and expressive.

What I don't like, though, is how he played fast and loose with the facts and the procedural history of the case, especially in the "fair use" context. I heard him say a variety of things that flew in the face of what I knew about this dispute that I've been following for a decade, longer than any other. But without verifying this based on the documents I've downloaded since August 2010, particularly by fact-checking against the Joint Appendix that was submitted to the Supreme Court (Volume 1 (pages 1-341), Volume 2 (pages 342-725)), I didn't want to accuse him of lying before at least refreshing my recollection. The bottom line is that his representations to the top U.S. court lacked veracity, even to the extent that I think he owes the justices an apology.

In the following, I'll highlight ten misrepresentations, in increasing order of relevance to the "fair use" decision:

10. "No summary judgment"

In connection with the standard of review, Oracle told the court that most fair use determinations are made by judges on summary judgment. This case is one, however, in which fair use was put before a jury, but the Federal Circuit determined that judgment as a matter of law (JMOL) was warranted. Justice Gorsuch asked Mr. Goldstein the following question:

JUSTICE GORSUCH: Briefly, just to follow up on -- on that, Justice Sotomayor's question.

Mr. Stewart [Deputy Solicitor General of the United States] argued that if -- if we were to uphold the jury verdict or send it back on fair use, that we would be negatively impacting summary judgment practice and that most district courts take these questions up as a matter of law in summary judgment.

Justice Gorsuch raised a legitimate concern, one that may very well be shared at least by his conservative colleagues, and Google will need at least one conservative vote to win a remand. As part of his response, Mr. Goldstein said: "Oracle didn't move for summary judgment in this case."

What Mr. Goldstein overlooks are three inconvenient facts, any single one of which makes his statement less than truthful:

  • In the first Federal Circuit appeal (the one that resulted in a copyrightability holding and a remand of "fair use" for retrial), Oracle clearly argued that the fair use question should be resolved by the judges and not be put before a jury.

  • On remand, Oracle simply wasn't allowed by Judge William H. "I taught myself Java" Alsup to bring a motion for summary judgment. Here's the related excerpt from a February 2, 2016 hearing transcript:

    THE [DISTRICT] COURT: All right. Now I want to go to another thing. I -- that you all are bombarding me with. I don't intend to entertain summary-judgment motions and dispositive motions. Whose idea was that?

    We had the Federal Circuit order me to give you a trial. I'm going to give the trial. And if the evidence shows that Rule 50 [judgment as a matter of law, which comes at a later stage than summary judgment] should be granted at the end of the trial, okay. I can do that on the evidence at trial.

  • Google itself had moved for summary judgment on fair use back in 2011 (and lost, as I reported at the time).

9. "Java SE was not in smartphones"

Google argues that its use of the Java API declaring code is "transformative," which would weigh in favor of fair use (the first factor is "the purpose and character of the use"). But the problem is that transformative use in copyright is clearest when you create, for instance, a parody of something, or a cover version of a song 50 years later that has a completely different style. Google just claims that they took Java SE (SE means "Standard Edition") to smartphones, but that has simply been disproven over and over.

While Mr. Goldstein conceded that Google didn't really have a strong case for the court determining that what Google did was fair use, so his realistic best case is a remand based on the standard of appellate review of a jury verdict, he nevertheless repeated that long-debunked misrepresentation to the Supreme Court:

"[T]he jury was entitled to conclude based on the record evidence that this was an entirely new context, the Java SE was not usable in this particular -- in a smartphone"

What a disgrace.

Even Danger, Andy Rubin's company that subsequently created Android, used Java SE in the T-Mobile Sidekick smartphone. On page 370 of the Joint Appendix I found that old Rubin testimony about how they used Java SE:

Q. And that was the Sidekick/Hiptop that we talked about?

A. Yes.

Q. And you put Java to SE APIs in Hiptop; is that right?

A. Yes. We created our own implementation of the Java 2 SE APIs for Hiptop.

11 pages before that passage, Mr. Rubin confirmed that the Sidekick was a smartphone:

Q. And when you described that Sidekick phone as one of the—the first smartphone, I believe you said, is it the kind of smartphone that we’re familiar with today, the modern Android and iPhones?

A. More or less. I mean, it did a lot of the same functionality. It allowed you to surf the Web, get the full Web on a phone. It had a larger screen. The screen could be in landscape or portrait mode. It did instant messaging. It did email and things like that.

So what Mr. Goldstein said was plain wrong even with respect to Java SE, in the words of Android's founder--and Java SE was furthermore used to build the SavaJe platform, "a Java OS for advanced mobile phones" as Wikipedia says. But focusing on just Java SE is nonsensically narrow anyway. Many years back I explained that Java itself was used in BlackBerry and Nokia smartphones.

8. "Google expert said expressly Android has not superseded Java SE"

This here relates to the fourth fair use factor ("the effect of the use upon the potential market for or value of the copyrighted work"). Oracle's multi-billion-dollar damages claim is based on how Android displaced Java in the smartphone market, where it already had significant traction as explained before.

The breadth of the following statement by Mr. Goldstein toward the end of the hearing makes it a falsehood or even something much worse:

"Mr. Rosenkranz says that Android supplanted and superseded Java SE. Page JA 255. The market harm expert says expressly Android has not superseded Java SE."

Actually, when I looked up the passage of the Joint Appendix that Mr. Goldstein, it turned out that he misleadingly took the statement out of context. There is one answer by Google's market harm expert that says the following:

"It has not superseded – Android has not superseded Java SE."

But just a few lines above, the same Google expert referred to Java SE as a programming platform for only personal computers! Here's the full passage, and it belies Mr. Goldstein's sweeping claim:

Q. Can you please tell us what your understanding is.

A. Java SE is one of the Java applications programming platforms. And it’s the one that specifically was designed for desktop computers.

Q. Did you reach an opinion as to whether or not Android had superseded Java SE in the market? [emphasis added]

A. I have.

Q. Can you please tell the jury what that opinion is.

A. It has not superseded – Android has not superseded Java SE.

Q. Do you have reasons for reaching that opinion? If so can, you explain them?

[1898] A. Yes, I have two reasons. The first is that the two products are on very different devices. As I just mentioned, Java SE is on personal computers. Android, on the other hand, is on smartphones. [emphasis added]

The full context leaves no room for doubt. But here, again, there's something to add: literally minutes after that Google expert left the stand, Google made an announcement (do you believe that timing was pure coincidence?) that Android was going to compete with Java SE in the desktop computer market (see two 2016 blog posts of mine about this: 1, 2).

7. "API declaring code is barely creative"

In connection with the second fair use factor ("the nature of the copyrighted work"), Mr. Goldstein made the following statement that flatly contradicts even Google's own testimony:

"importantly, the fact that the original material here, the declarations, is barely creative"

As I already reported in 2011, Joshua Bloch, Google's chief Java engineer and top Java expert, testified the following:

Q. Would you say that designing APIs is a creative activity?

[objection to form, by Google lawyer]

THE WITNESS: Yes, absolutely.

So here you have Google's Supreme Court lawyer in 2020 saying the opposite of what Google's top in-house Java expert said nine years earlier when he testified under oath. Which one of the two do you believe? And which one of the two (if any) would you say is more likely to have told a lie than the other?

6. "new declarations only written in a new language"

Throughout this dispute Google has been trying to portray the APIs as being inextricably linked to the "free" Java programming language (the commands). Even the district court didn't buy that.

But Mr. Goldstein reiterated this on Wednesday:

"The computer scientists' brief at page 18, the Microsoft brief at 14, explain that both Apple and Microsoft, Oracle's examples, did re-implement prior interfaces. The reason that they didn't use these interfaces is they were using a different language, as if they were writing in French, rather than English."

As I noted yesterday, the Supreme Court is well aware of the fact that Apple and Microsoft (the latter even supporting Google here on fair use, though not on copyrightability) created smartphone operating systems without stealing anything from Oracle. Sure, Apple relied initially on Objective C and Microsoft on C# (the latter being pretty basically a Java clone, which I personally used a lot and like a great deal). But Oracle's Supreme Court brief notes that Spring and Log4J both wrote different declaring code for their own prewritten programs in the Java language that perform similar functions to those in Java SE.

5. "APIs never licensed separately from the Java language"

This is technically somewhat related to the previous item. And just as untruthful:

"The evidence at trial, for example JA 56, is the former CEO of Oracle saying that the APIs were never licensed or sold separately from the language, in contrast to his just base [?] assertion that IBM was paying for it."

Sun and Oracle offered the Specification license, which does exactly that, and large companies like IBM, Microsoft, SAP, Red Hat, and Oracle (before it acquired Sun) all took Specification licenses to the declaring code and then wrote their own implementing code. The Joint Appendix mentions all of that, mostly on pages 301-304 and 402-409. It's a fact that I fought against Oracle's acquisition of Sun, especially but not only in the EU, and Microsoft and SAP were co-complainants, which is also well-documented in the media. They wanted to prevent the deal from happening because of that very type of license!

Also, don't be misled: the "CEO of Oracle" here was Jonathan Schwartz, who despite his legal department telling him as early as 2007 that he should sue Google, never wanted to act. He was Sun's last CEO, he failed, and that's why Sun lost its independence. Oracle had to buy Sun because it was about to go bankrupt without an acquirer. Again, I know that subject well because I was a vocal opponent to the deal. I even organized and conducted a Wall Street analyst briefing (Westin on Times Square) in October 2009.

4. "no proof IBM paid for declaring code"

The passage I quoted under the previous subhead suggests that IBM was not paying for it. But there was undisputed trial testimony that IBM, and others, paid for a license to the declaring code without simultaneously licensing the implementing code (pages 301-304 and 402-410 of the Joint Appendix).

3. "Oracle seeking to block millions of apps"

Mr. Goldstein said:

"Here you have minimally creative declarations and they are being invoked to block the publication of millions of programs on an innovative smartphone platform."

The "minimally creative" part, which contradicts Google's own in-house Java expert's testimony, is totally incorrect anyway, but what's just as wrong is this allegation that the case is about blocking millions of apps in any way.

Oracle gives all programmers a free license to use Java SE to write apps that run on numerous licensed platform implementations of Java SE, including from Oracle's competitors. I'm sure I'd know if Oracle had ever sought any kind of relief against a single app or app developer just because they used the Java APIs. This case has nothing to do with app developers writing applications that run on Java SE. This case is about companies who build rival platforms. Oracle is simply seeking to require companies who seek to use Oracle’s investment to create a direct rival platform to take a license and agree to the compatibility requirements that benefit developers. All rival platform developers complied--until Google decided to just use the stuff without taking either an open-source or a commercial license, both of which Sun offered.

2. "industry expectations and practice"

Industry doesn't make the law. Congress does, courts interpret it. But courts do take the world outside into consideration, and Mr. Goldstein said the following:

"I don't think there is actual debate about the expectations of the industry. And they have nothing to do with licensed reuse of interfaces. The --there is a widespread consensus in the industry and among computer scientists that this has been the practice."

Concern about "upending" an industry practice is a potential factor in the Supreme Court's decision. But what Google's lawyer describes as "undisputed" was actually controverted by some, and even the district court disagreed with Google. It excluded Google's "industry custom" expert because he did not have the goods. He sought to testify about an industry custom of unlicensed copying without identifying any examples of unlicensed copying — relying instead on licensed copying (page 470). Having failed to establish the record for its argument in the most sympathetic court it ever faced in its history, Google tried to establish the point with amici curiae who similarly did not distinguish between licensed and unlicensed re-implementations and who conflated open-source use pursuant to a license with unlicensed use. At any rate, large software companies like SAS, Synopsys, and Mathworks along with the CEO of EMC (Joseph Tucci) and the former CEO and CTO of Sun Microsystems (Scott McNealy) all submitted amicus briefs refuting that there was such a thing as an industry-wide consensus of unlicensed copying being above board.

1. "benefit to developers"

It's almost funny how Google tries to paint a picture of being so altruistic that it really just had the best in mind for app developers ("reusing the minimally creative declarations allowed the developers to write millions of creative applications that are used by more than a billion people"--and "minimally creative" was addressed further above, but here the focus is on the alleged concern for developers).

It's not that I, as an app developer, am not grateful to both Apple (which didn't take anything from Oracle/Sun) and Google for the market opportunity they've created. My posts on the ongoing app store antitrust matters (such as 1 and 2) hopefully reflect my efforts to be a voice of reason in that context amid all the bashing of platform makers that I see elsewhere on the web.

But as a developer I depend on copyright and also want platform makers to respect it. Google could have benefited developers without Java, and if it wanted to do Java, it could simply have taken any of multiple licenses offered by Sun and Oracle and benefited developers. Google even rejected the free open-source license. All that Google would have had to do was agree that it and anyone who reused its code would also make their code available on such terms ("copyleft"). Google refused that quid pro quo because it concluded it would not be in its business interest (Joint Appendix, page 367). Helping programmers looks like a pretext to me, also in light of the various ways that Google protects its own APIs now and tries to lock app programmers into Android in various ways.

The bottom line is that Mr. Goldstein said a number of things that were misleading at best and mendacious at worst. He's impressive, but he may have gone too far in his vigorous efforts to secure the best possible outcome for his client Google.

Share with other professionals via LinkedIn: