Saturday, June 1, 2013

EFF's amicus briefs in Oracle v. Google suggest non-copyrightability is only path to interoperability

Many roads lead to Rome, but if all you have is a hammer, everything looks like a nail. I'm sure you've heard these two proverbs before, but the authors and signatories of two EFF-orchestrated amicus curiae briefs supporting Google against Oracle's Android-Java copyright appeal did not bear them in mind when making their submissions to the Federal Circuit, proposing to throw out the baby with the bath water by denying copyrightability to some of the most creative and expressive elements of computer programs for the sake of (whatever they mean by) interoperability.

In my previous post I explained that Google and the EFF's advocacy of "interoperability" is anything but principled, and that Android is not interoperable with Java by any stretch of the imagination. But for the purposes of this present post I'm going to forget about Android's incompatibility with Java and focus on the question of how to ensure true interoperability while affording software creations a reasonable degree of intellectual property protection.

In most cases software copyright isn't an obstacle to interoperability because the creators of application programming interfaces (APIs) frequently have an interest in the adoption of their creations. Self-regulation works, and it actually is possible to do clean-room implementations of the Java APIs, subject to reasonable terms and conditions that Google didn't want to accept. But every now and then there could be a problem with a right holder wielding copyright in an effort to prevent interoperability, just like there are standard-essential patent (SEP) owners, like Google, who pursue injunctions. What do we do then?

If true interoperability is at stake, the "fair use" exception might work. The EFF's "Start-ups" brief mentions "fair use" only twice, in the same paragraph, in an effort to portray "fair use" cases (Sony, Sega) as non-copyrightability decisions. Its "Scientists" brief doesn't mention it even once.

Even if "fair use" failed in a particular case to ensure interoperability (for example, it's specific to copyright and doesn't apply to patents), there would always be antitrust law as a last resort in order to obtain a compulsory license on FRAND terms. The EFF's "Start-ups" brief doesn't mention "antitrust" or "competition law" at all. Its "Scientists" brief refers to it only once, in connection with UNIX, but fails to address it otherwise, even in connection with IBM's conduct in the mainframe business and the effect it had on its PC-related strategies.

Both briefs presuppose that there is only one tool: copyrightability. But besides the non-copyrightability hammer, there are those other tools I just discussed. All it takes the appeals court to disagree with Google's lawyers' and amigos' anti-copyrightability argument is to agree with Google's in-house copyright counsel William Paltry, who advocated that the "appropriate balance" can best be struck with the help of "fair use" and other tools. The whole notion of an interoperability-copyrightability conflict goes away once there is a determination that interoperability is irrelevant to copyrightability and better resolved with the help of more flexible and use-specific tools. Copyrightability should be consistent from the moment of creation, just like patent offices don't reject applications that could result in standard-essential patents because of their subject matter (and patent law doesn't even have the "fair use" tool). Addressing protection first, access later means that creators don't become the victims of their own creations' success, while still being able to take market power and "expectations" into consideration, with the benefit of 20/20 hindsight, when someone wants to use someone else's creation. But some people want the Federal Circuit to destroy intellectual property with a hammer only to deprive itself and all other courts in the United States of the opportunity to decide on interoperability based on what happens to an API after its creation and on what a defendant wants to do with it and to it.

The EFF's submissions are, of course, consistent with Google's appellate brief, which even argued that intellectual property protection can be lost over time, mentioning Aspirin as one example. Not only are trademarks, contrary to copyright, a totally use-dependent intellectual property right but Google's lawyers omit a historic fact. A couple of readers contacted me on Twitter shortly after the post I just linked to and noted that Bayer lost trademark protection for Aspirin because it had to give up certain rights under the Treaty of Versailles (after World War I). The most diplomatic term for this is expropriation.

I'm now going to explain why there are (at least) two EFF amicus briefs, not just one, and then I'll quote some passages from those briefs and comment on them.

Two EFF amicus curiae briefs: one submitted in the name of "Computer Scientists", the other in the name of "Innovators, Start-ups, and Investors"

On Thursday the EFF issued a press release on the first of its amicus briefs in this case, which the EFF submitted as counsel to a group named "Computer Scientists". The EFF also published this brief. I have the greatest respect for these amici, which the EFF listed on its website with short bios, and for the EFF attorneys representing them. But for the reasons I explained in a previous post, all those Google-supporting amici (the ones on that list and all others) fail to counterbalance the economic weight of those supporting Oracle, and the EFF used FUD to mobilize people and companies for its cause. It says something that most amici only signed up as private persons because they couldn't get their current or former employers to support Google on this one.

The second EFF-orchestrated brief, published near-simultaneously on academic document site SSRN, is positioned as a submission by "Software Innovators, Start-ups, and Investors" and was filed by the Law, Technology, and Public Policy Clinic of the EFF's Vice Chairwoman of the Board, Professor Pamela Samuelson. On its website the Samuelson Clinic lists the EFF as its first client. One of the Computer Scientists' counsel is Professor Jason Schultz, who coauthored the Defensive Patent License with his colleague, Assistant Professor Jennifer Urban. Professor Schultz also was an Assistant Professor at the Samuelson Clinic while the DPL was originally conceived. The "Software Innovators, Start-ups, and Investors" include two former EFF leaders -- co-founder Mitch Kapor and former Chairwoman of the Board (1991-1998; the EFF was founded in 1990) Esther Dyson -- and Esther Dyson's EDventure Holdings. (The "Computer Scientists" also include several EFF board members: John Perry Barlow, Brian Behlendorf, David Farber).

Far be it from me to be disrespectful of Professor Samuelson and her staff. And I'm not accusing the EFF of circumventing any filing limitations or of finding ways to get more speaking time at the appellate hearing. I just want to shed light on the fact that support for Google is not nearly as broadbased as the EFF would have us believe. This effort is orchestrated, not organic.

Oracle's amici represent, among other things, a substantial percentage of software R&D, and include some of Oracle's fiercest competitors, while Google has not gathered support from any of its rivals.

The signatories of the EFF's briefs don't have a comparably high share among computer scientists, software innovators, start-ups, and investors as Oracle's corporate amici have in the overall software market. For example, there are only two venture funds among the investors the EFF mobilized: Foundry Group and Union Square Ventures. Those aren't California-based, which is where most of the venture capital action in the U.S. is (Sand Hill Road). They're also great, but only two, and about 90% of the anti-software-patent commentary that you'll find in the media (including social media) from venture investors comes from partners of these funds (Brad Feld and Fred Wilson), who are apparently more anti-IP than I thought (and than they probably are, but it appears to be easy to mobilize them for this kind of agenda).

The two EFF briefs are the only ones to have been published so far.

Commentary on select passages from the EFF briefs

The EFF's "Computer Scientists" brief talks a lot about the IBM PC and its BIOS (Basic Input/Output System) as an example of an open API and speculates that "IBM enforced copyright on the BIOS source code and would certainly have used copyright to control reimplementations of the BIOS API if the law allowed". But there's a much better explanation for IBM's decision. From the 1950s on (and even at the beginning of this decade) IBM had antitrust problems with its mainframe monopoly. While mainframes are mentioned in that same brief as one of various industries allegedly made innovative and competitive by "[t]he freedom to reimplement and extend existing APIs", and an AT&T consent decree is mentioned in connection with UNIX, the 1956 IBM Consent Decree is omitted. It's no less famous than the AT&T one, but less supportive of the EFF's argument. It shows that interoperability can be achieved by antitrust means. And when IBM created the PC, it had learned its lesson well and tried to steer clear of running into the same problems again in the next computing market. So it allowed the creation of IBM PC compatibles -- basically, in order to obviate the need for another consent decree.

One of the "Computer Scientists" is Andrew Tridgell, who created Samba, which the brief also discusses. Again, the antitrust approach worked, and Microsoft was required by the European Commission to offer a FRAND license to all relevant intellectual property (patents, copyright, whatever). The EFF already made a Samba-based anti-copyrightability argument last year, but it fails to explain why non-copyrightability is the answer (and "fair use" and FRAND are not).

Some of the EFF's examples are protocols that are far simpler than the Java APIs in question. For example, HTTP or the menu structure at issue in the Lotus case. I'm not saying that there's no copyright in simpler things than the Java APIs, but there's definitely a stronger case for non-copyrightability in cases like HTTP or a spreadsheet menu than in the Java case.

The EFF's briefs also suggest that software is cheaper if APIs aren't copyrightable. For example, the EFF gives an example of how NASA allegedly saved money because it was able to continue to use some old manufacturing robots even though the relevant APIs were "orphan works" ("The freedom to reimplement APIs also helps rescue 'orphan' software or data—systems whose creators have either gone out of business or abandoned their product in the marketplace."). The EFF then links this to budget cuts and the latest sequester. This "logic" runs counter to the positions taken very recently by the Federal Circuit's Chief Judge Rader in a snowplow patent case: "While the general public certainly enjoys lower prices, cheap copies of patented inventions have the effect of inhibiting innovation and incentive. This detrimental effect, coupled with the public’s general interest in the judicial protection of property rights in inventive technology, outweighs any interest the public has in purchasing cheaper infringing products. In sum, the public has a greater interest in acquiring new technology through the protections provided by the Patent Act than it has in buying 'cheaper knock-offs.'"

The EFF's "Innovators, Start-ups, and Investors" brief says a lot of things about the value of interoperability that I wholeheartedly agree with, beginning with the first paragraph:

"Interoperability between programs and systems is key to software innovation. It allows systems to connect and individuals to apply knowledge and skills from one environment to another. By easing the way for software developers to build upon existing platforms, interoperability allows efficient software ecosystems to grow, fueling the development of innovative new products and services and increasing competition to the benefit of consumers."

It's also true that "[c]loud services rely heavily on interoperability--and specifically on APIs--to work". But neither EFF brief explains why the denial of copyrightability is the only way to ensure interoperability. There's some talk about legal certainty, but if "fair use" was too uncertain to be useful, it wouldn't be as important as it is. FRAND is also given more and more meaning by courts (despite Google's efforts to derail or delay some of the related processes).

What I strongly dislike about the "Innovators, Start-ups, and Investors" brief are various references to the "use" of APIs. That's not at issue in this case. Oracle is happy to let anyone "use" the Java APIs to write Java apps.

The other key point I want to criticize about it is the lack of balance when it comes to the interests of start-ups in connection with intellectual property protection. Professor Samuelson herself has conducted studies such as the Berkeley Patent Survey that show, for example, that start-ups with patents are more likely to be funded by venture capitalists. It's simply wrong -- and inconsistent with the Samuelson Clinic's findings in other contexts -- to suggest that weaker IP, merely by lowering entrance barriers in some areas, is the perfect environment for startups to flourish. If every incumbent could force a startup out of the market by disallowing any use of APIs, there would be a problem, but that's not what would result from a reversal of the district court's non-copyrightability decision. There's another way to look at it: if even Oracle couldn't prevent Google from hijacking Java, what would a startup do if Google decided to supplant its products or services? IP is quite often needed to give startups leverage against incumbents.

The right balance must be struck, and in my opinion Google and the EFF's positions are too far on the anti-IP side. That's why I can't support them, though I support the cause of interoperability (more consistently than they do).

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: