For the masochistic, Groklaw has compiled the expert reports in an accessible fashion here and here. Why do I look at the reports, and not the briefs? It turns out that lawyers will make all sorts of arguments about what the evidence will say, but what is really relevant is the evidence actually presented. The expert reports, submitted before trial, are the broadest form of evidence that can be admitted – the court can whittle down what the jury hears, but typically experts are not allowed to go much beyond their reports.

These reports represent the best evidentiary presentation the parties have on the technical merits. It turns out that as a factual matter, both reports overlap quite a bit, and neither seems “wrong” as a matter of technical fact. I would sure hope so – these are pretty well respected professors and, quite frankly, the issues in this case are just not that complicated from a coding standpoint. (Note: for those wonder what gives me the authority to say that, I could say a lot, but I’ll just note that in a prior life I wrote a book about software programming for an electronic mail API).

What level of abstraction was presented and argued to the jury? As far as I can tell from the reports, other than a couple or three routines that were directly copied, the Oracle’s expert found little or no similar structures or sequences in the main body source code – the part that actually does the work. The only similarity – and it was nearly identical – was in the structure, sequence and organization of the grouping of function names, and the “packages” or files that they were located in.

For computer nerds, also identical were function names, parameter orders, and variable structures passed in as parameters. In other words, the header files were essentially identical. And they would have to be, if the goal is to have a compatible system. The inputs (the function names and parameters) and the outputs need to be the same. The only way you can disallow this usage of the API is to say that you cannot create an independent software program (even one of this size) that mimics the inputs and outputs of the original program.

To say that would be bad policy, and as I discuss below, probably not in accordance with precedent. This is why the experts are both right. Oracle’s expert says they are identical, and Google copied because that was the best way to lure application developers – by providing compatibility (and the jury agreed, as to the copying part). Google’s expert says, so what? The only thing copied was functional, and that’s legal. It’s this last part that a) led to the hung jury, and b) the court will have to rule on.

In my last post, I assumed that the level of abstraction must have been at a deeper level than just the names of the methods. Why did I do that?

First, the court’s jury instructions make clear that function names are not at issue. But I guess the court left it to the jury whether the collection could be infringed.

Second, the idea that an API could be infringed is usually something courts decide well in advance of trial, and it’s a question that doesn’t usually make it to trial.

Third, based on media accounts, it appeared that there was more testimony about deeper similarities in the code. The copied functions, I argued in my prior post, supported that view. Except that there were no other similarities. I think it is a testament to Oracle’s lawyers (and experts) that this misperception of a dirty clean room shone through in media reports, because the actual evidence belies the media accounts.

This is why I decided to dig deeper, and why one should not rely on second hand reports of important evidence. Based on my reading of the reports (and I admit that I could be missing something – I wasn’t in the courtroom), I think that the court will have no choice but to hold that the collection of API names is uncopyrightable – at least at this level of abstraction and claimed infringement.

To the extent that there are bits of non-functional code, I would say that’s probably fair use as a matter of law to implement a compatible system. I made a very similar argument in an article I wrote 12 years ago – long before I went into academia.

Prof. Boyden asked in a comment to my prior post whether there was any law that supported the copying of APIs structure and header files. I think there is: Lotus v. Borland. That case is famous for allowing Borland to mimic the Lotus structure, but there was also an API of sorts. Lotus macros were based on the menu structure, and to provide program compatiblity with Lotus, Borland implemented the same structure. So, for example, in Lotus, a user would hit “/” to bring up the menus, “F” to bring up the file menu, and “O” to bring up the open menu. As a result, the macro “/FO” would mimic this, to bring up the open menu.

Borland’s product would “read” macro programs written for Lotus, and perform the same operation. No underlying similarity of the computer code, but an identical API that took the same inputs to create the same output the user expected.

Like the lower court here, the lower court there found infringement of the structure, sequence, and organization of the menu structure. Like the lower court here, the court there found it irrelevant that Borland got the menu structure from third-party books rather than Lotus’s own product. (Here, Google asserts that it got the API’s from Apache Harmony, a compatible Java system, rather than the Java documents themselves). There is some dispute about whether Sun sanctioned the Apache project, and what effect that should have on the case. I think that the Harmony is a red herring.The reality is that it does not matter either way – a copy is a copy is a copy – if the copy is illicit that is.

In Lotus, the lower court found the API creative and copyrightable, the very question facing the court here. On appeal, however, the First Circuit ruled that the API was a method of operation, likening it to the buttons on a VCR. I think that’s a bit simplistic, but it was definitely the right ruling. The case went up to the Supreme Court, and it was a blockbuster case, expected to — once and for all — put this question to rest.

Alas, the Supreme Court affirmed without opinion by an evenly divided court. And the circuit court ruling stood. And it still stands – the court never took another case, and the gist of Lotus v. Borland has been applied over and over, but rarely as directly as it might apply here.

Wholesale, direct compatibility copying of APIs just doesn’t happen very often, and certainly not on the scale and with the stakes of that at issue here. Perhaps that is why there is no definitive case holding that an entire API structure is uncopyrightable. You would think we would have by 2012, but nope. Lotus comes close, but it is not identical. In Lotus, the menu structure was much smaller, and the names and structure were far less creative. Further, the concern was macro programming written by users for internal use that would not allow them to switch to a new spreadsheet program. Java programs, on the other hand, are designed to be distributed to the public in most cases.

Then again, the core issue is the same: the ability to switch the underlying program while maintaining compatibility of programs that have already been written. Based on this similarity, my prediction is that Judge Alsup will say that the collection of names is not copyrightable, or at the very least usage of the API in this manner is fair use as a matter of law. We’ll see if I’m right, and whether an appeals court affirms it.

Here’s the short story of what happened, as far as I can gather.

1. Google needed an application platform for its Android phones. This platform allows software developers to write programs (or “apps” in mobile device lingo) that will run on the phone.

2. Google decided that Sun’s (now Oracle’s) Java was the best way to go.

3. Google didn’t want to pay Sun for a license to a “virtual machine” that would run on Android phones.

4. Google developed its own virtual machine that is compatible with the Java programming language. To do so, Google had to make “APIs” that were compatible with Java. These APIs are essentially modules that provide functionality on the phone based on a keywords (instructions) from a Java language computer program. For example, if I want to display “Hello World” on the phone screen, I need only call print(”Hello World”). The API module has a bunch of hidden functionality that takes “Hello World” and sends it out to the display on the screen – manipulating memory, manipulating the display, etc.

5. The key dispute is just how much of the Java source code was copied, if any to create the Google version.

The jury today held the following:

1. One small routine (9 lines) was copied directly – line for line. The court said no damages for this, but this finding will be relevant later

2. Google copied the “structure, sequence, and organization” of 37 Java API modules. I’ll discuss what this means later.

3. There was no finding on whether the copying was fair use – the jury deadlocked.

4. Google did not copy any “documentation” including comments in the source code.

5. Google was not fooled into thinking it had a license from Sun.

To understand any of this, one must understand the levels of abstraction in computer code. Some options are as follows:

A. Line by line copying of the entire source code.

B. Line by line paraphrasing of the source code (changing variable names, for example, but otherwise idential lines).

C. Copying of the structure, sequence and organization of the source code – deciding what functions to include or not, creative ways to implement them, creative ways to solve problems, creative ways to name and structure variables, etc.  (The creativity can’t be based on functionality)

D. Copying of the functionality, but not the stucture, sequence and organization – you usually find this with reverse engineering or independent development

E. Copying of just the names of functions with similar functionality – the structure and sequence is the same, but only as far as the names go (like print, save, etc.). The Court ruled already that this is not protected.

F. Completely different functionality, including different structure, sequence, organization, names, and functionality.

Obviously F was out if Google wanted to maintain compatibility with the Java programming language (which is not copyrightable).

So, Google set up what is often called a “cleanroom.” The idea is not new – AMD famously set up a cleanroom to develop copyrighted aspects of its x86 compatible microprocessors back in the early 1990’s. Like Google now (according to the jury), AMD famously failed to keep its cleanroom clean.

Here’s how a cleanroom works. One group develops a specification of functionality for each of the API function names (which are, remember, not protected – people are allowed to make compatible programs using the same names, like print and save). Ideally, you do this through reverse engineering, but arguably it can be done by reading copyrighted specifications/manuals, and extracting the functionality. Quite frankly, you could probably use the original documentation as well, but it does not appear as “clean” when you do so.

Then, a second group takes the “pure functionality” description, and writes its own implementation. If it is done properly, you find no overlapping source code or comments, and no overlapping structure, sequence and organization. If there happens to be similar structure, sequence and organization, then the cleanroom still wins, because that similarity must have been dictated by functionality. After all, the whole point of the cleanroom is that the people writing the software could not copy because they did not have the original to copy from.

So, where did it all go wrong? There were a few smoking guns that the jury might have latched on to:

1. Google had some emails early on that said there was no way to duplicate the functionality, and thus Google should just take a license.

2. Some of the code (specifically, the 9 lines) were copied directly. While not big in itself, it makes one wonder how clean the team was.

3. The head of development noted in an email that it was a problem for the cleanroom people to have had Sun experience, but some apparently did.

4.  Oracle’s expert testified (I believe) that some of the similarities were not based on functionality, or were so close as to have been copied. Google’s expert, of course, said the opposite, and the jury made its choice. It probably didn’t help Google that Oracle’s expert came from hometown Stanford, while Google’s came from far-away Duke.

So, the jury may have just discounted the Google cleanroom story, and believed Oracle’s. And that’s what it found. As someone who litigated many copyight cases between competing companies, this is not a shocking outcome. This issue will not doubt bring the copyright v. functionality issue to the forefront (as it did in Lotus v. Borland and Intel v. AMD), this stuff is bread and butter for most technology copyright lawyers. It’s almost always factually determined. Only the scope of this case is different in my book – everything else looks like many cases I’ve litigated (and a couple that I’ve tried).

So, what happens now in the copyright phase?  (A trial on patent infringement started today.) Judge Alsup has two important decisions to make.

First, the court has to decide what to do with the fair use ruling. Many say that a mistrial is warranted since fair use is a question of fact and the jury deadlocked. I’m not so sure. The facts on fair use are not really disputed here – only the legal interpretation of them; my experience is that courts are more than willing to make a ruling one way or the other when copying is clear (as the jury now says it is). I don’t know what the court will do, but my gut says no fair use here.  My experience is that failed cleanrooms fail fair use – it means that what was copied was more than pure functionality, and it is for commercial use with market substitution. The only real basis for fair use is that the material copied was pure functionality, and that’s the next inquiry.

Second, the court must determine whether the structure, sequence, and organization of these APIs can be copyrightable, or whether they are pure functionality. I don’t know the answer to that question. It will depend in large part on:

a. whether the structure, etc., copied was at a high level (e.g. structure of functions) or at a low level (e.g. line by line and function by function);

b. the volume of copied (something like 11,000 lines is at issue);

c. the credibility of the experts in testifying to how much of structure that is similar is functionally based.  On a related note, the folks over at groklaw think for the most part think this is not copyrightable. They have had tremendous coverage of this case.

I’ve been on both sides of this argument, and I’ve seen it go both ways, so I don’t have any predictions. I do look forward to seeing the outcome, though. It has been a while since I’ve written about copyright law and computer software; this case makes me want to rejoin the fray.

Now compare

Content should slip the TV versus computer snag soon. So, hello, Stage 3; thank you commodity cloud computing.

Customers want their content on demand. College students forgo cable fees, because they are so damn expensive and carry mainly crap they don’t want. Streams work for them, because they have the campus network. But many I know have cut the cable and gone to streaming only. And why not? Lower cost is clear. Plus, no one talks about whatever is today’s Friends at the water cooler because a #1 show is nowhere near as watched. Must see T.V.? Please.

I will bet that the demand for direct delivery of content will mean a new order for T.V. and film. As a technical matter, Ed Felten reminded me that asynchronous delivery of content poses problems. He knows far more than I. But consumers will want to buy (or direct subscribe to) content and will use the Internet to get that content. Producers like HBO will lead the way save for threats from cable companies. Assuming tensions in that sphere, someone will figure out how to leverage current network advances to store content cheaply, deliver it so that peaks are handled, and cable boxes will go the way of the DoDo.

In other words, why rely on cable for the menu of content? Cable’s value is the delivery of whatever content someone wants. The odd part is that I still subscribe to cable and will even watch a movie with commercials (Ocean’s 11 on TNT, Star Wars) when it is one rather than getting the DVD I own. But my way will die. I may abandon it too. And if the older folks stray, look out cable. The young ‘uns are already gone. More will follow. Then again the Lakers are on tonight. So maybe sports will save cable. Then again, sports teams own their cable stations when possible. Hey cable, say it with me: NBA, NFL, MLB, NHL on demand with a little ESPN for kicks?

I’ve got several points I want to make:

1. The overall outcome is not a surprise, to me at least.
2. On the issue of “What, beyond responding to individual takedown notices, do ISPs have to do to maintain their Section 512 immunity?” the answer is a) something — which is a win for plaintiffs — but b) not much, which is a win for defendants.
3. The court I think identifies the correct role for “red flags” — at least, it was the one argued for in the amicus brief I drafted — but on the question of what constitutes “knowledge” I think the opinion is less successful. The distinction drawn between “specific vs. general” knowledge, both here and in connection with secondary liability generally, is I think not helpful and in any event will soon break down.
4. The court helpfully ducks the massive red herring, subject of a considerable amount of argument in the briefs, blog commentary, and mentioned in the opinion below, over whether the fact that some allegedly infringing works may in fact be authorized should somehow bear on the interpretation of Section 512.
5. It’s interesting and perhaps unsurprising that the court avoids legislative history and purpose as tools for interpreting the statute — with one significant (and I think salutary) exception.
6. Just like the exclusionary rule sometimes warps Fourth Amendment criminal procedure law, there is reason here to be concerned that statutory damages might warp the substantive law governing the liability of ISPs.
7. I have some higher-level thoughts on what the decision means for content owners and ISPs.

That’s a pretty long list, actually, so I’m going to have to break this up into multiple posts.

The Outcome Is Not a Surprise

The district court opinion essentially read portions out of the statute. The district court held that, even though there was evidence presented that YouTube executives knew of some infringing activity, thus raising at least a disputed issue of material fact as to the “knowledge or awareness” exceptions in Section 512(c)(1)(A), nevertheless the safe harbor applied because “when YouTube was given [takedown] notices, it removed the material. It is thus protected ‘from liability for all monetary relief for direct, vicarious and contributory infringement’ subject to the specific provisions of the DMCA.” 718 F. Supp. 2d at 526. The “knowledge or awareness” exceptions applied only to “knowledge of specific and identifiable infringements of particular individual items” — which the court appeared to equate with “the DMCA notification regime.” Id. at 523-24. In other words, the only way an ISP can get the requisite knowledge under Section 512(c)(1)(A) is to receive a takedown notice specifically identifying an infringing file by URL — which of course is exactly what Section 512(c)(1)(C) provides for (and Section 512(c)(3)(B)(i) makes clear that a defective notice — a letter from a content owner that does not comply with Section 512(c)(3) — does not give the requisite knowledge, so that is not the purpose of the “knowledge or awareness” exceptions).

Second, the court also essentially read out the “vicarious-liability-like” exception in Section 512(c)(1)(B) (more about that later), which says that an ISP gets no immunity if it “does not receive a financial benefit directly attributable to the infringing activity, in a case in which the service provider has the right and ability to control such activity.” The district court inexplicably added a “knowledge” element to this, even though knowledge is neither part of the traditional vicarious liability concept nor mentioned in the statute, and held that where an ISP did not specifically know of a particular infringing file, its ability to control such activity was irrelevant. As the Second Circuit noted, this also had the effect of collapsing (c)(1)(B) onto (c)(1)(A), which already deals with “knowledge” situations.

The end result was that the district court held that the only evidence that was material in deciding whether YouTube fell within the Section 512(c) safe harbor was whether it complied with takedown notices. But given that the plaintiffs were not even challenging YouTube’s response to takedown notices, the district court’s decision essentially made all of the discovery in the case irrelevant — it could have been written three years earlier on a motion to dismiss. Indeed, that is probably what explains the fact that the district court issued its opinion a mere nine days after briefing on the cross-motions for summary judgement, which had taken three months to complete, were fully submitted, without even so much as holding oral argument. Some indication of the novelty of the district court’s holding I think is indicated by the fact that YouTube did not even file a motion to dismiss, but instead proceeded directly to an answer. (I wonder if lawyers at YouTube were kicking themselves for that decision after reading the district court’s opinion.)

It turns out that that was all too easy. The Second Circuit decision unsurprisingly held that the nullified provisions of Section 512(c) actually must mean something. Thus, Judge Cabranes, writing the opinion, referred repeatedly to the well-established canon that “statutory interpretations that render language superfluous are disfavored.” Section 512(c)(1)(A) excludes immunity for ISPs that know of infringing activity or are aware of facts or circumstances from which infringing activity is apparent; Section 512(c)(1)(B) excludes immunity where an ISP has the “right and ability to control” the activity and directly profits from it. The Second Circuit remanded for further consideration of whether either of these exceptions was satisfied (given the first go-round, I’m not sanguine about avoiding another appeal).

But that still leaves the question of what, precisely, the Second Circuit thinks these exceptions mean. More on that in the next post.

Appeal from the judgment of the United States District Court for the Southern District of New York (Louis L. Stanton, Judge), granting summary judgment to the defendants-appellees on all claims of direct and secondary copyright infringement based on a finding that the defendants-appellees were entitled to safe harbor protection under the Digital Millennium Copyright Act (“DMCA”), 17 U.S.C. § 512. Although the District Court correctly held that the § 512(c) safe harbor requires knowledge or awareness of specific infringing activity, we vacate the order granting summary judgment because a reasonable jury could find that YouTube had actual knowledge or awareness of specific infringing activity on its website. We further hold that the District Court erred by interpreting the “right and ability to control” infringing activity to require “item-specific” knowledge. Finally, we affirm the District Court’s holding that three of the challenged YouTube software functions fall within the safe harbor for infringement that occurs “by reason of” storage at the direction of the user, and remand for further fact-finding with respect to a fourth software function.

Affirmed in part, vacated in part, and remanded.

The book is described here (OUP site) and here (Amazon). The Introduction and Table of Contents are available here.

Short abstract:

“Infrastructure resources are at the center of many contentious public policy debates, ranging from what to do about our crumbling roads and bridges, to whether and how to protect of our natural environment, to patent law reform, to electromagnetic spectrum allocation, to providing universal health care, to energy policy, to network neutrality regulation and the future of the Internet. Each involves a battle to control infrastructure resources, set the terms and conditions under which the public gets access, and determine how the infrastructure and various infrastructure-dependent systems evolve over time. This book advances strong economic arguments for managing and sustaining infrastructure resources as commons. The book identifies resource valuation and attendant management problems that recur across many different fields and many different resource types, and it develops a functional economic approach to understanding and analyzing these problems and potential solutions.”

Other fun things about the song came up while I poked around. I thought that the style and voice reminded me of something or someone, namely Sting. The voice and the lyrics “have your friends collect your records and then change your number” thorws me back to “I see you’ve sent my letters back and my LP records and they’re all scratched”. It turns out that others agree (per the great Wikipedia). (another video below Gotye vs. The Police apparently thought so too). The Goyte video (below) also has throw backs to Peter Gabriel’s breakthrough videos of the 1980s (like Sledgehammer and Big Time), and Gotye has been compared to Gabriel as well. Throw on top of all that source or inspiration a connection to Ba Ba Black Sheep which connects to Twinkle Twinkle little star (circa 1780 for the Mozart tune) (yet another video parodying Gotye and its source NOTE: this one is a little coarse) and holy copyinspirationcreation Batman! So on a rainysunny weekend, enjoy the music and try not to parse who, what, where, how copyinspirationcreation happens. Just revel.

PS the female singer, Kimbra is great. I pasted one of her videos below too. It’s called Settle Down. More on her in another post. Again just enjoy for now.

Darth Vader? Good. Bag Pipes? Good. Darth and bagpipes put together with a guy in a kilt and on a unicycle? I leave that judgment to posterity (and you all).

Print is Dead. Long Live the Word. Britannica Stops the Presses. Welcome to the Henry Blake cliche festival. CNN Money reports that after 244 years the print edition of Britannica will no longer be offered. As many may recall, one study indicated the Wikipedia was more close to as accurate than Britannica. It may come as no surprise to those who know me that I tend to ask questions. My parents we of the “Look it up” school of thought. They bought World Book (remember them?) which fell short of my needs quickly. Then they bought a set of Britannica. It was lovely. Leather (or simulated perhaps), gilt edged, the micro and macro pedias, lined up in the den on wooden shelves. I ceased talking to my parents and went to the books. I loved them. In grade school, I learned that they were not to be cited but used to guide deeper research. Yes, grade school. So I was quite fortunate. My parents could afford such a luxury, and I reveled in it.

But let’s not obsess over print. Yes, analog copies are more difficult to reach out and destroy. I questioned the ability to manipulate e-books when I wrote “One possibility of the new technology is that books will continually evolve as authors change their mind or update a text. This idea brings images of revisionist Greedo shootings.” That ability was connected to Orwell in theory, and then when Amazon in fact used the power to remove a book, (remember it was in fact Orwell’s 1984?). But think about the costs for buying the research tool that was a multi-volume set. Today the print edition is $1395. I think was more when I was a kid, and that they used a model familiar to academics and software users (pay for updates) to generate revenue after the first sale. You also had to have room for the books. Digital divide and access to knowledge discussions can miss that the cost of the set would cover Internet access for 20 months. Of course one needs a computer too. But the computer and the Internet access can do much more than access one set of data. I suppose someone could study the cost of paper, binding, and shipping compared to the energy and materials for a computer and connection to see the true saving or lack of it. I will bet the numbers favor general purpose tech (Frischmann infrastructure ideas may be invoked here).

Digital also is a dream for the look it up model. I disagree with Carr and the Shallows analysis here. Yes, I look up things when on my e-reader (still a Kindle in fact) or online. And guess what I return to the text. I taunt students when they fail to look up words or ideas despite having the Internet at their fingertips almost all the time. To me online resources are great and to be embraced while also addressing the archiving and other issues new technologies raise.

Britannica’s President Jorge Cauz said some interesting and funny things to print junkies, “Everyone will want to call this the end of an era, and I understand that,” Cauz says. “But there’s no sad moment for us. I think outsiders are more nostalgic about the books than I am.” Given that he stated print was “less than 1% of the company’s total sales” he seems wise, and one wonders at why they didn’t kill it sooner. Other curious facts include that the online version is only 15% of revenue and “The other 85% is sales of education products: online learning tools, curriculum products and more.”

Will folks pay for the online version at $70 per year? I would guess not. Nonetheless Cauz claims that people interested in expert opinions will turn to Britannica: “Google’s algorithm doesn’t know what’s fact or what’s fiction,” Cauz concedes. “So Wikipedia is often the No. 1 or No. 2 result on search. But I’d bet a lot of money that most people would rather use Britannica than Wikipedia.” So far the evidence seems to be to the contrary. Wikipedia seems to hold up well. Stanford’s Encyclopedia of Philosophy is great too. I have argued that commons-based, Benkler goods could collapse, but for now they seem to be doing well.

So go with God, Britannica. Thank you for the years of service and enhancing my childhood. And congratulations on your new form. Like those in Good To Great, you have ditched the old method and seek to play in the new space. It is a bet, but it you are in the correct game and that is good.