In some regards, including key topics of patent reform, Google may be the best friend among large U.S. tech companies that software developers can find. Those of you who've been following this blog over these past four years may have noted a steady increase in support for Google's positions in connection with patent policy and patent disputes (though some areas of disagreement, such as FRAND, remain) and even copyright (Google Books). But with respect to software copyright, I'm starting to doubt whether I'll ever agree with Google in my lifetime because things have just gotten a whole lot worse.
The off-base reasoning that Google had persuaded Judge Alsup to adopt in the Oracle case prompted a former U.S. copyright chief, Ralph Oman, to issue a warning that this line of thought "largely eviscerates copyright protection for some of the most creative aspects of computer software." After the United States Court of Appeals for the Federal Circuit exposed the district court ruling as grossly erroneous, Google has apparently altered course and is trying a new trick in its continuing defense against Oracle's copyright infringement claim. While I understand that a party's argument has to be ever more strategic as a case goes higher, there's no written or unwritten requirement to become more radical. Unfortunately, what Google is trying to achieve has gone from "largely eviscerat[ing] copyright protection for some of the most creative aspects of computer software" (emphasis mine) to totally eliminating copyright protection for all computer software. I've been doing this for long enough to know that many people will find it hard to believe this portrayal of Google's approach (in fact, if I read something like this on another site, I'd also want to know why). That's why I needed to think about it a little more after my initial reaction to Google's cert petition. But it's actually not hard to explain now.
The question to begin with is this:
How would Google's proposed question for the Supreme Court have to be answered in order for Google to achieve its desired reversal of fortune?
Let's look at that proposed question again:
"Whether copyright protection extends to all elements of an original work of computer software, including a system or method of operation, that an author could have written in more than one way."
In my initial reaction I already highlighted that the word "written" omits the fact that only expressive material is protected, and not everything that is written is expressive. Since copyright was not initially invented for the purpose of protecting software (nor were patents, but they were at least meant to protect applied science from the outset), there's a dichotomy: copyright protects the expressive aspects of program code, but not its functionality. Still, copyright has served software developers very well (for decades already) in all cases in which outright copying occurred. It would most likely also work well in cases in which, for example, someone translates program code from one programming language to another. What it's not designed for is to prevent independent implementations (including intentional imitations as well as inadvertent ones) of someone else's problem-solving strategies. That's what patents are for. That's why I have filed patent applications for three inventions this year as opposed to just relying on copyright.
Oracle v. Google is a case of outright copying. Google never claimed that it came up with 7,000 lines of API declaring code from Java independently. It admits that it took the stuff, and argues that whatever it took was for the taking.
This gets us to the fundamental reason for which Google's objectives in the Oracle case (unlike Google's policy in certain other areas) run counter to the interests of software developers large and small. In order to get away with what it's done and wants to continue to do (instead of reaching a commercial agreement with Oracle, Sun's acquirer and thus the rightful owner of the material at issue), Google has only one choice: it must convince the courts that even outright copying of substantial amounts of program code is permitted by copyright law. If this sounds paradoxical, that's because it is.
Google is, now more than ever, trying to move the line between copyrightable expression and non-copyrightable functionality so far into and even beyond the range of highly-creative, expressive code that copyright no longer applies.
It's just that I don't want to sound alarmist. Otherwise I would (and based on Google's proposition I could) say that Google is advocating the death of software copyright and, as a result, the death of commercial software except for areas in which trade secret and trademark protection may be sufficient. The reason I don't want to say so is simply that I don't believe Google will make much headway, if any. That said, if someone tried to kill me by throwing a few small cotton balls at me, I wouldn't be afraid but I'd still be shocked at the intent.
As Boston-based tech and IP lawyer Lee Gesmer mentioned (in the only opinion on Google's Supreme Court petition I've read or heard so far), "the Supreme Court accepts review of approximately 1% of cases appealed to it." There are reasons to assume that this case has a better chance than 1% of being heard. In any event, the question Google raised is one that Oracle's lead counsel on appeal, Orrick's Joshua Rosenkranz, was eager to address at the Federal Circuit hearing last December, where he began his oral argument in the following way:
"Your Honors, in this copyright case, where extraordinary creativity is conceded, the central question is whether the creative computer code loses all copyright protection under Section 102(b) just because it also serves a function."
He didn't get to continue because a circuit judge immediately changed the focus--Mr. Rosenkranz would just have preached to the converted on this one. Without (ever) knowing how this oral argument would have continued without the quick interruption, it's obvious that Oracle's counsel wouldn't have raised this question as the very first one if he had not been confident of his ability to win the case on this basis.
Let's get back to the question I outlined further above: How would Google's question for the Supreme Court have to be answered in order for Google to fend off Oracle's claim? And what would be the further consequences of a Google win for everyone including a little app developer like me?
"Whether copyright protection extends to all elements of an original work of computer software, including a system or method of operation, that an author could have written in more than one way."
Besides blurring the distinction between what's written and what's expressive, this question implies two things, either one of which and especially the combination of which is an all-out attack on software copyright:
The "including" part has a clear answer in statutory law: according to Section 102(b), copyright protection does not extend to a system or method of operation. Raising this question would be of use to Google only if the Supreme Court concluded that the Federal Circuit afforded copyright protection against the statute. The suggested "No" answer to this part would, however, require the question Oracle's counsel outlined at the beginning of his Federal Circuit argument to be answered as follows: "Yes, even extraordinarily creative computer program code loses all copyright protection just because it also serves a function."
The dichotomy between expressive and functional aspects of program code would go away. The 7,000 lines of highly-creative declaring code would not be protected just because they also serve a function.
Now that I've thought about it a little more, I see a dangerous radicalization of Google's position in this line of thought. Previously, Google made a misguided (but initially-successful) attempt to advocate an interoperability exception to copyrightability (which was also wrong because Android isn't Java-compatible). Now it is no longer limiting its argument to program code that plays a role in interoperability. It's now suggesting that by affording copyright protection to 7,000 lines of highly-creative code, the Federal Circuit gave Oracle copyright (against the statute) over technical functionality.
The "more than one way" part is equally problematic. While the answer is easy (the creativity hurdle is very low under U.S. copyright law), "more than one way" may be an appropriate description with respect to isolated lines of the 7,000-line material at issue, and even then only with respect to the simplest ones of those lines. My favorite amicus brief submitted to the Federal Circuit in connection with this case explained the creative choices in API design very well. It gave an example of how even a circle-drawing function could be designed in multiple ways. But if we assume, as an example and only for the sake of the argument, that there is an average of only three (way too low for many of the 7,000 lines!) alternative ways to express the underlying ideas of each of the 7,000 lines, there would be 3 to the power of 7,000 alternative ways. I entered that number into a spreadsheet and the number was way out of range. (Even if Google had taken only 70 -- 1% of 7,000 -- lines of declaring API code with three alternatives per line, the number of possible combinations would be a number with more than 30 zeroes.) And this formula doesn't even include the possibility of having different functions assigned to different classes.
If the Supreme Court, in order to agree with Google, held that Sun Microsystems' engineers didn't actually have a sufficient number of alternative ways to express the ideas underlying those 7,000 lines of Java API code, no software would be copyrightable anymore.
Google is now way past the point at which any advocacy group can back its positions in the name of software developers large and small. Even those software developers who would like API-related code not to be copyrightable should reject Google's latest approach because, in the alternative, they would attempt commercial suicide only to help one big company in a dispute with another major player.
I would also encourage my fellow software developers to think about what Google's attempt to make all software code non-copyrightable if it's functional means for the protection of object code. This case here is about source code. If Oracle lost in a case in which source code is copied, none of us could prevail in a case in which object code is stolen. For example, the guys who stole the Flappy Bird object code and just replaced the image of the bird would be legally above board.
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:
