Insanity Wins As Appeals Court Overturns Google's Fair Use Victory For Java APIs

Oh, CAFC. The Court of Appeals for the Federal Circuit has spent decades fucking up patent law, and now they're doing their damndest to fuck up copyright law as well. In case you'd forgotten, the big case between Oracle and Google over whether or not Google infringed on Oracle's copyrights is still going on -- and it appears it will still be going on for quite a while longer, as CAFC this morning came down with a laughably stupid opinion, overturning the district court's jury verdict, which had said that Google's use of a few parts of Java's API was protected by fair use. That jury verdict was kind of silly in the first place, because the whole trial (the second one in the case) made little sense, as basically everyone outside of Oracle and the CAFC had previously understood (correctly) that APIs are simply not covered by copyright.

Section 102(b) of the Copyright Act says quite clearly:

In no case does copyright protection for an original work of authorship extend to any idea, procedure, process, system, method of operation, concept, principle, or discovery, regardless of the form in which it is described, explained, illustrated, or embodied in such work.

And an API is pretty clearly a procedure, process, system or method of operation -- it's just instructions on how to access certain elements, similar to a recipe. But, CAFC (who shouldn't be hearing this case in the first place) simply couldn't be bothered to comprehend what an API is, and insisted that because to the judges non-technical brains, an API looks the same as software, it must be copyrightable as software. That was after Judge Alsup in the district court (following trial #1) had spent quite a lot of time explaining to CAFC why APIs are not copyrightable.

Thus, we had the second trial, which was weird, because all of the arguments about fair use, were couched in this weird "um, well, it's not really copyrightable at all, but CAFC says it's copyrightable, so let's just say it's fair use" argument. And the jury then said, yes, it's fair use.

But CAFC has now rejected that and sent the case back to the lower court for a third trial for damages. And, while we normally expect bad reasoning from CAFC decisions, this one is particularly stupid. In short, CAFC's reasoning is basically "we think this is infringement, and thus we're going to handwave around the law to make sure that it's infringement." It's bad. And, again, CAFC shouldn't even be hearing the case. CAFC hears appeals on patent cases, and originally there were a few patent claims in this case, but they all got dumped at a very early stage. So this case should have gone to the 9th Circuit (who also might have messed it up, but it has at least a marginally better record than CAFC).

Anyway, the CAFC does an awful lot of handwaving around historical precedent to justify its decision to basically start from scratch in going through the fair use four factors. As we've discussed multiple times, one of the problems with the four factors test is that it allows a court to choose who it likes better, and then twist the four factors to give it the outcome it wants. That appears to be what is happening here. On the first factor (nature of the use), CAFC basically says "we think the jury is stupid, this is obviously commercial use, and thus it goes against Google." Google had argued (and the jury had implicitly agreed) that because Google doesn't charge for Android, and interoperability and progress of innovation enabled by using similar API setups are not inherently commercial motives, that this was fair use. The court basically says "but Google has so much money!" which is not how a fair use argument works. But... CAFC.

That Google might also have non-commercial motives is irrelevant as a matter of law. As the Supreme Courtmade clear when The Nation magazine published excerptsfrom Harper & Row's book, partly for the purpose ofproviding the public newsworthy information, the question"is not whether the sole motive of the use is monetarygain but whether the user stands to profit from exploitationof the copyrighted material without paying the customaryprice." Harper & Row, 471 U.S. at 562. Second,although Google maintains that its revenue flows fromadvertisements, not from Android, commerciality does notdepend on how Google earns its money. Indeed, "[d]irecteconomic benefit is not required to demonstrate a commercialuse." A&M Records, 239 F.3d at 1015. We find,therefore, that, to the extent we must assume the juryfound Google's use of the API packages to be anythingother than overwhelmingly commercial, that conclusionfinds no substantial evidentiary support in the record.Accordingly, Google's commercial use of the API packagesweighs against a finding of fair use.

On the question of transformative use, the lower court had suggested that using the APIs in a "fresh context" (such as for smartphones) could be seen as transformative. But, again CAFC says "nope."

Google's arguments are without merit. As explainedbelow, Google's use of the API packages is not transformativeas a matter of law because: (1) it does not fit withinthe uses listed in the preamble to 107; (2) the purpose ofthe API packages in Android is the same as the purpose ofthe packages in the Java platform; (3) Google made noalteration to the expressive content or message of thecopyrighted material; and (4) smartphones were not anew context.

CAFC also, once again, shows that it still doesn't understand why APIs and code are not the same thing:

That Google wrote its own implementing code is irrelevantto the question of whether use of the APIs wastransformative. As we noted in the prior appeal, "noplagiarist can excuse the wrong by showing how much ofhis work he did not pirate." Oracle, 750 F.3d at 1375(quoting Harper & Row, 471 U.S. at 565). The relevantquestion is whether Google altered "the expressive contentor message of the original work" that it copied-notwhether it rewrote the portions it did not copy.

But that's not the relevant question. The relevant question is how the copied text is being used, and in Google's case it's in a completely different context, which is why it had to write its own implementing code. Again, this is fallout from CAFC's earlier wrong decision in that it still does not understand that API instructions are not the same as implementing code itself.

The court also more or less laughs at the idea that Google had "good faith" reasons for doing what it did:

Ultimately, we find that, even assuming the jury wasunpersuaded that Google acted in bad faith, the highlycommercial and non-transformative nature of the usestrongly support the conclusion that the first factorweighs against a finding of fair use.

On factor two of the four factors analysis -- the nature of the work -- CAFC again actually gives that one to Google and says that the jury could have found that weighed towards fair use, but immediately points out that it's going to give this factor less weight, because... it still doesn't understand the difference between APIs and software, and insists that if it gave this factor any weight, it would destroy the idea of copyright for software. Of course, that's only true if you can't tell the difference between an API and software.

We note, moreover, that allowing this one factorto dictate a conclusion of fair use in all cases involvingcopying of software could effectively negate Congress'sexpress declaration-continuing unchanged for some fortyyears-that software is copyrightable. Accordingly,though the jury's assumed view of the nature of thecopyrighted work weighs in favor of finding fair use, it hasless significance to the overall analysis.

On factor three, concerning the amount use, the court again is so incredibly confused it's frustrating. Google has long held (and the jury and the lower court appeared to agree) that it used the bare minimum of the Java API to enable basic interoperability in building apps for Android. But CAFC is so tied up in its own made up reality, that it insists no reasonable jury could agree with this:

Even assuming the jury accepted Google's argumentthat it copied only a small portion of Java, no reasonablejury could conclude that what was copied was qualitativelyinsignificant, particularly when the material copiedwas important to the creation of the Android platform.

Eventually, though CAFC says factor 3 could be either "neutral" or "arguably weighs against" fair use.

On to factor four -- the impact on the market. Factors one and four tend to be the ones that courts focus on the most. And, here, Oracle has spent plenty of time whining that its own total failure to capitalize on Java in the mobile market should be blamed on Google's copying piece of its API as opposed to the reality, which is that Oracle completely botched its own efforts in the space, because it's Oracle. The court more or less ignores various points about how Oracle wasn't really in the mobile market and more or less says "Google was super successful, so clearly it was at the expense of Oracle."

Given the record evidence of actual and potentialharm, we conclude that "unrestricted and widespreadconduct of the sort engaged in by" Google would result in"a substantially adverse impact on the potential marketfor the original" and its derivatives.

And, thus, the court then weighs factors one and four most heavily, and says it's not fair use, and back to Northern California we go for trial number three, specifically on the damages question. I imagine Google will appeal this, either asking for an en banc rehearing or to the Supreme Court. The Supreme Court refused to hear the previous request on the question of copyrightability of APIs, so this is probably a long shot. However, there are a few oddities within the ruling that maybe will catch someone's attention at the Supreme Court (and the Supreme Court has spent years dunking on CAFC and mocking its dumb decisions, so maybe they'd like to do that again).

Honestly, the most concerning part of the whole thing is how much of a mess CAFC has made of the whole process. The court ruled correctly originally that APIs are not subject to copyright. CAFC threw that out and ordered the court to have a jury determine the fair use question. The jury found it to be fair use, and even though CAFC had ordered the issue be heard by a jury, it now says "meh, we disagree with the jury." That's... bizarre.

Anyway, considering the importance of interoperability in software, this case is yet another potential disaster, limited only by the fact that CAFC doesn't cover any particular region and most copyright cases aren't influenced by CAFC precedent since those cases would flow up through the other appeals courts. However, you can see how someone wishing to go after others for API infringement might just throw in a silly patent claim just get it into CAFC. Hopefully the Supreme Court steps in and fixes this, but if not, it would be nice (I know, I know...) if Congress actually stepped up and pointed out that APIs are not covered by copyright at all. And while they're at it, they should burn CAFC to the ground. It was a dumb idea in the first place and should be shut down.