Sunday, May 8, 2016

Oracle v. Google copyright retrial won't bring clarification on application programming interfaces (APIs)

Tomorrow, the Oracle v. Google Android-Java copyright retrial is scheduled to begin. Almost six years have passed since the filing of this lawsuit, and about four years since the first trial, which could have been much more useful if not for Judge Alsup's sometimes unfathomable (and bad) decisions.

There's really no reason to get excited about this retrial with respect to application programming interfaces (APIs). For Oracle it's more important than for Google to make headway, and for the outside world it's of very limited interest what happens now (as opposed to what may happen on appeal). During and after the first trial, the copyrightability of APIs was a key issue. In my opinion, the way this played out was merely consistent with what the law had been all along, but admittedly a lot of people took a different position in the public debate, so this had to be settled--and it has been, in Oracle's (and almost all software developers') favor.

Nothing that the judge, the jury or the lawyers will say during the retrial is going to have an impact on API copyrightability or result in a general rule that makes APIs available on "fair use" terms. The issues that will be raised are predictable based on pretrial motions practice and pretrial orders. There'll be some well-known industry execs testifying (pretty much the same lineup as last time), they'll show some company-internal documents, and they'll discuss mindblowing financials. But let me put it in the clearest terms:

This retrial is not about whether the technically-incompatible reimplementation of an API is covered by fair use.

This is not a prediction of the outcome. While I'm as convinced as ever that there is hardly a clearer case of UNfair use than this one (let me just refer you to two previous blog posts on this question: "Fair use" is a fairy tale, Google Books comparison), the trial is a tossup. If I had to estimate the chances of Google persuading the jury that this is (though it absolutely positively isn't) "fair use," or of Google getting another jury to deadlock (like in 2012, where some jurors made comments to the press that didn't suggest they understood the case), then I'd say it's a little bit more likely than Oracle convincing the jury to throw out Google's defenses. In numbers: I see a 55% or 60% chance for Google. But if Google prevails, Oracle can still win by judicial decision. Anything is possible, even a post-trial finding by Judge Alsup that no reasonable jury could have found "fair use" here, though he couldn't surprise me more (not because of the merits, which would warrant such a decision, but because of the way he's been handling the case so far). The appeals court, however, would be fairly likely side with Oracle.

I saw a comment somewhere a few days ago from a lawyer who said that copyright trials like this favor right holders because the jury, after seeing literal copying, will be more inclined to conclude that there was an infringement. One can only make that statement without conducting an analysis of the record of this particular case. Google will benefit from the following success factors (which are not going to matter much on appeal, but they will influence the jury):

  1. Jury instructions: Judge Alsup has made some edits since his first draft, which was scandalous. Those jury instructions still suck. There are issues on which a clear holding in Oracle's favor would have been warranted. Instead, the final jury instructions (still) suggest that you can steal someone's program code as long as the product resulting from such theft is in some way, shape or form different from the original. It's a shame.

  2. Bifurcation: The same judge who had no qualms about letting an entire jury trial over "fair use" be a potential waste (including long jury deliberations) instead of previously making up his mind on copyrightability suddenly (honi soit qui mal y pense) found it more economic to separate the damages question from the "fair use" merits. The net effect is that Google will actually be able to present some evidence and testimony during the first trial stage that the jury could easily confuse for "fair use" considerations, or that might at least influence the jury for purely psychological reasons. By contrast, Oracle's ability to raise some of its most powerful points with respect to Google's conduct has been limited.

  3. "Free" Java language: Unlike in almost any other "fair use" case, Oracle's lawyers face a considerable hurdle (not the only one, but a key one) in the fact that Sun allowed everyone to implement the Java language. While it will (based on pretrial filings and orders) be made technically clear to the jury that only a few APIs were deemed necessary for implementation of the Java language, jurors may very well feel that Java was for the taking anyway and that whatever Google took beyond the undisputedly-free part was then covered by "fair use." Jonathan Schwartz, on whose 2012 testimony I commented in this blog post, will be in the tank for Google once more with a particular emphasis on "free Java."

  4. Open source: I've been dealing with open source IP issues ever since I became involved with MySQL AB, a dual-licensing open-source startup, almost 15 years ago. To me it's just obvious that Google simply didn't use Java on open-source terms in the past; it may do so in the future, but that's another story. Java under the GPL and Java in Android (Apache-licensed) have nothing to do with each other, and whatever the Harmony project may have done with out a license doesn't mean Google's actions were legal. However, it won't be easy for jurors to see through the OpenJDK, GPL, Apache, Harmony etc. smokescreen they're going to have in front of them.

Based on what I've been able to research, Oracle's trial lawyers know how to build a connection with a jury and how to make compelling, simple arguments on issues that are actually very complex. So do Google's. I expect both trial teams to be at a level. And Google's team has the advantages outlined in the previous four bullet points, and possibly others.

If the jury is very smart, it will figure this out and conclude that there was no "fair use" here, by far. Of course, a stupid jury could also reach that conclusion, but more likely a stupid or even "semi-smart" jury will side with Google.

If the jury agrees with Oracle, Google will appeal. Otherwise, Oracle will. But this "fair use" thing here should never even have been put before a jury. The evidence simply doesn't support a reasonable finding of "fair use." Even one or more of the "fair use" cases Google cited in the appellate proceedings involved a determination by judges that they, not juries, could resolve "fair use" since it's an equitable rule, as opposed to a mere fact-finding exercise.

No doubt both parties will try to get a judgment as a matter of law (JMOL). I doubt they'll get it. The correct decision would be to enter JMOL in Oracle's favor, but it's too far a stretch of the imagination that Judge Alsup would do that. Chances of that one are not nil, but maybe 0.1%. It's also unlikely he'll side with Google because he probably knows this could result in a humiliating reversal, which would call his fairness into question considering that the appeals court already found the 2012 decision in Google's favor ridiculous. The safest path for him will be to let the trial unfold and to let the jury render a verdict (or deadlock).

There's simply no way in which the outside world, especially the industry at large, will be smarter after this trial. In tactical terms, either party obviously would increase its chances in the appellate proceedings, at least psychologically, by prevailing now. But it will only get interesting at the appeals court(s). And even then, this is not really going to be about APIs. There's particularly one passage in the jury instructions (relating to how hard or easy it is to establish transformative use) that I think raises huge issues for the software industry at large, but that one doesn't mention or allude to APIs in any way.

So if anybody thought this trial could effectively provide everyone with free access to APIs, including technically-incompatible reimplementations, you can only be disappointed. Even if Google wins, that won't be the result. The result would, in that case, just be that a jury got confused about open-source licensing issues, "free" access to a programming language and things like that.

If you'd like to be updated on the smartphone patent disputes and other intellectual property matters I cover, please subscribe to my RSS feed (in the right-hand column) and/or follow me on Twitter @FOSSpatents and Google+.

Share with other professionals via LinkedIn: