Jun 13, 2021
The largest global power during the rise of intellectual property was England, so the world adopted her philosophies. The US had the same impact on software law.
Most case law that shaped the software industry is based on copyright law. Our first real software laws appeared in the 1970s and now have 50 years of jurisprudence to help guide us. This episode looks at the laws, supreme court cases, and some circuit appeals cases that shaped the software industry.
In our previous episode we went through a brief review of how the modern intellectual property laws came to be. Patent laws flowed from inventors in Venice in the 1400s, royals gave privileges to own a monopoly to inventors throughout the rest of Europe over the next couple of centuries, transferred to panels and academies during and after the Age of Revolutions, and slowly matured for each industry as technology progressed.
Copyright laws formed similarly, although they were a little behind patent laws due to the fact that they weren’t really necessary until we got the printing press. But when it came to data on a device, we had a case in 1908 we covered in the previous episode that led Congress to enact the 1909 Copyright Act.
Mechanical music boxes evolved into mechanical forms of data storage and computing evolved from mechanical to digital. Following World War II there was an explosion in new technologies, with those in computing funded heavily by US government. Or at least, until we got ourselves tangled up in a very unpopular asymmetrical war in Vietnam. The Mansfield Amendment of 1969, was a small bill in the 1970 Military Authorization Act that ended the US military from funding research that didn’t have a direct relationship to a specific military function. Money could still flow from ARPA into a program like the ARPAnet because we wanted to keep those missiles flying in case of nuclear war. But over time the impact was that a lot of those dollars the military had pumped into computing to help develop the underlying basic sciences behind things like radar and digital computing was about to dry up. This is a turning point: it was time to take the computing industry commercial. And that means lawyers.
And so we got the first laws pertaining to software shortly after the software industry emerged from more and more custom requirements for these mainframes and then minicomputers and the growing collection of computer programmers. The Copyright Act of 1976 was the first major overhaul to the copyright laws since the 1909 Copyright Act. Since then, the US had become a true world power and much as the rest of the world followed the British laws from the Statute of Anne in 1709 as a template for copyright protections, the world looked on as the US developed their laws. Many nations had joined the Berne Convention for international copyright protections, but the publishing industry had exploded. We had magazines, so many newspapers, so many book publishers. And we had this whole weird new thing to deal with: software.
Congress didn’t explicitly protect software in the Copyright Act of 1976. But did add cards and tape as mediums and Congress knew this was an exploding new thing that would work itself out in the courts if they didn’t step in. And of course executives from the new software industry were asking their representatives to get in front of things rather than have the unpredictable courts adjudicate a weird copyright mess in places where technology meets copy protection. So in section 117, Congress appointed the National Commission on New Technological Uses of Copyrighted Works, or CONTU) to provide a report about software and added a placeholder in the act that empaneled them.
CONTU held hearings. They went beyond just software as there was another newish technology changing the world: photocopying. They presented their findings in 1978 and recommended we define a computer program as a set of statements or instructions to be used directly or indirectly in a computer in order to bring about a certain result. They also recommended that copies be allowed if required to use the program and that those be destroyed when the user no longer has rights to the software. This is important because this is an era where we could write software into memory or start installing compiled code onto a computer and then hand the media used to install it off to someone else.
At the time the hobbyist industry was just about to evolve into the PC industry, but hard disks were years out for most of those machines. It was all about floppies. But up-market there was all kinds of storage and the righting was on the wall about what was about to come. Install software onto a computer, copy and sell the disk, move on. People would of course do that, but not legally.
Companies could still sign away their copyright protections as part of a sales agreement but the right to copy was under the creator’s control. But things like End User License Agreements were still far away. Imagine how ludicrous the idea that a piece of software if a piece of software went bad that it could put a company out of business in the 1970s. That would come as we needed to protect liability and not just restrict the right to copy to those who, well, had the right to do so. Further, we hadn’t yet standardized on computer languages. And yet companies were building complicated logic to automate business and needed to be able to adapt works for other computers and so congress looked to provide that right at the direction of CONTU as well, if only to the company doing the customizations and not allowing the software to then be resold. These were all hashed out and put into law in 1980.
And that’s an important moment as suddenly the party who owned a copy was the rightful owner of a piece of software. Many of the provisions read as though we were dealing with book sellers selling a copy of a book, not dealing with the intricate details of the technology, but with technology those can change so quickly and those who make laws aren’t exactly technologists, so that’s to be expected.
Source code versus compiled code also got tested. In 1982 Williams Electronics v Artic International explored a video game that was in a ROM (which is how games were distributed before disks and cassette tapes. Here, the Third Circuit weighed in on whether if the ROM was built into the machine, if it could be copied as it was utilitarian and therefore not covered under copyright. The source code was protected but what about what amounts to compiled code sitting on the ROM. They of course found that it was indeed protected.
They again weighed in on Apple v Franklin in 1983. Here, Franklin Computer was cloning Apple computers and claimed it couldn’t clone the computer without copying what was in the ROMs, which at the time was a remedial version of what we think of as an operating system today. Franklin claimed the OS was in fact a process or method of operation and Apple claimed it was novel. At the time the OS was converted to a binary language at runtime and that object code was a task called AppleSoft but it was still a program and thus still protected. One and two years later respectively, we got Mac OS 1 and Windows 1.
1986 saw Whelan Associates v Jaslow. Here, Elaine Whelan created a management system for a dental lab on the IBM Series One, in EDL. That was a minicomputer and when the personal computer came along she sued Jaslow because he took a BASIC version to market for the PC. He argued it was a different language and the set of commands was therefore different. But the programs looked structurally similar. She won, as while some literal elements were the same, “the copyrights of computer programs can be infringed even absent copying of the literal elements of the program.” This is where it’s simple to identify literal copying of software code when it’s done verbatim but difficult to identify non-literal copyright infringement.
But this was all professional software. What about those silly video games all the kids wanted? Well, Atari applied for a copyright for one of their games, Breakout. Here, Register of Copyrights, Ralph Oman chose not to Register the copyright. And so Atari sued, winning in the appeal.
There were certainly other dental management packages on the market at the time. But the court found that “copyrights do not protect ideas – only expressions of ideas.” Many found fault with the decision and the Second Circuit heard Computer Associates v Altai in 1992. Here, the court applied a three-step test of Abstraction-Filtration-Comparison to determine how similar products were and held that Altai's rewritten code did not meet the necessary requirements for copyright infringement.
There were other types of litigation surrounding the emerging digital sphere at the time as well. The Computer Fraud and Abuse Act came along in 1986 and would be amended in 89, 94, 96, and 2001. Here, a number of criminal offenses were defined - not copyright but they have come up to criminalize activities that should have otherwise been copyright cases. And the Copyright Act of 1976 along with the CONTU findings were amended to cover the rental market came to be (much as happened with VHS tapes and Congress established provisions to cover that in 1990. Keep in mind that time sharing was just ending by then but we could rent video games over dial-up and of course VHS rentals were huge at the time.
Here’s a fun one, Atari infringed on Nintendo’s copyright by claiming they were a defendant in a case and applying to the Copyright Office to get a copy of the 10NES program so they could actually infringe on their copyright. They tried to claim they couldn’t infringe because they couldn’t make games unless they reverse engineered the systems. Atari lost that one. But Sega won a similar one soon thereafter because playing more games on a Sega was fair use. Sony tried to sue Connectix in a similar case where you booted the PlayStation console using a BIOS provided by Connectix. And again, that was reverse engineering for the sake of fair use of a PlayStation people payed for. Kinda’ like jailbreaking an iPhone, right? Yup, apps that help jailbreak, like Cydia, are legal on an iPhone. But Apple moves the cheese so much in terms of what’s required to make it work so far that it’s a bigger pain to jailbreak than it’s worth. Much better than suing everyone.
Laws are created and then refined in the courts. MAI Systems Corp. v. Peak Computer made it to the Ninth Circuit Court of Appeals in 1993. This involved Eric Francis leaving MAI and joining Peak. He then loaded MAI’s diagnostics tools onto computers. MAI thought they should have a license per computer, but yet Peak used the same disk in multiple computers. The crucial change here was that the copy made, while ephemeral, was decided to be a copy of the software and so violated the copyright. We said we’d bring up that EULA though. In 1996, the Seventh Circuit found in ProCD v Zeidenberg, that the license preempted copyright thus allowing companies to use either copyright law or a license when seeking damages and giving lawyers yet another reason to answer any and all questions with “it depends.”
One thing was certain, the digital world was coming fast in those Clinton years. I mean, the White House would have a Gopher page and Yahoo! would be on display at his second inauguration. So in 1998 we got the Digital Millennium Copyright Act (DMCA). Here, Congress added to Section 117 to allow for software copies if the software was required for maintenance of a computer. And yet software was still just a set of statements, like instructions in a book, that led the computer to a given result. The DMCA did have provisions to provide treatment to content providers and e-commerce providers. It also implemented two international treaties and provided remedies for anti-circumvention of copy-prevention systems since by then cracking was becoming a bigger thing. There was more packed in here. We got MAI Systems v Peak Computer reversed by law, refinement to how the Copyright Office works, modernizing audio and movie rights, and provisions to facilitate distance education. And of course the DMCA protected boat hull designs because, you know, might as well cram some stuff into a digital copyright act.
In addition to the cases we covered earlier, we had Mazer v Stein, Dymow v Bolton, and even Computer Associates v Altai, which cemented the AFC method as the means most courts determine copyright protection as it extends to non-literal components such as dialogue and images. Time and time again, courts have weighed in on what fair use is because the boundaries are constantly shifting, in part due to technology, but also in part due to shifting business models.
One of those shifting business models was ripping songs and movies. RealDVD got sued by the MPAA for allowing people to rip DVDs. YouTube would later get sued by Viacom but courts found no punitive damages could be awarded. Still, many online portals started to scan for and filter out works they could know were copy protected, especially given the rise of machine learning to aid in the process. But those were big, major companies at the time. IO Group, Inc sued Veoh for uploaded video content and the judge found Veoh was protected by safe harbor.
Safe Harbor mostly refers to the Online Copyright Infringement Liability Limitation Act, or OCILLA for short, which shields online portals and internet service providers from copyright infringement. This would be separate from Section 230, which protects those same organizations from being sued for 3rd party content uploaded on their sites. That’s the law Trump wanted overturned during his final year in office but given that the EU has Directive 2000/31/EC, Australia has the Defamation Act of 2005, Italy has the Electronic Commerce Directive 2000, and lots of other countries like England and Germany have had courts find similarly, it is now part of being an Internet company. Although the future of “big tech” cases (and the damage many claim is being done to democracy) may find it refined or limited.
In 2016, Cisco sued Arista for allegedly copying the command line interfaces to manage switches. Cisco lost but had claimed more than $300 million in damages. Here, the existing Cisco command structure allowed Arista to recruit seasoned Cisco administrators to the cause. Cisco had done the mental modeling to evolve those commands for decades and it seemed like those commands would have been their intellectual property. But, Arista hadn’t copied the code.
Then in 2017, in ZeniMax vs Oculus, ZeniMax wan a half billion dollar case against Oculus for copying their software architecture.
And we continue to struggle with what copyright means as far as code goes. Just in 2021, the Supreme Court ruled in Google v Oracle America that using application programming interfaces (APIs) including representative source code can be transformative and fall within fair use, though did not rule if such APIs are copyrightable. I’m sure the CP/M team, who once practically owned the operating system market would have something to say about that after Microsoft swooped in with and recreated much of the work they had done. But that’s for another episode.
And traditional media cases continue. ABS Entertainment vs CBS looked at whether digitally remastering works extended copyright. BMG vs Cox Communications challenged peer-to-peer file-sharing in safe harbor cases (not to mention the whole Napster testifying before congress thing). You certainly can’t resell mp3 files the way you could drop off a few dozen CDs at Tower Records, right? Capitol Records vs ReDigi said nope. Perfect 10 v Amazon, Goldman v Breitbart, and so many more cases continued to narrow down who and how audio, images, text, and other works could have the right to copy restricted by creators. But sometimes it’s confusing. Dr. Seuss vs ComicMix found that merging Star Trek and “Oh, the Places You’ll Go” was enough transformativeness to break the copyright of Dr Seuss, or was that the Fair Use Doctrine? Sometimes I find conflicting lines in opinions. Speaking of conflict…
Is the government immune from copyright? Allen v Cooper, Governor of North Carolina made it to the Supreme Court, where they applied blanket copyright protections. Now, this was a shipwreck case but extended to digital works and the Supreme Court seemed to begrudgingly find for the state, and looked to a law as remedy rather than awarding damages. In other words, the “digital Blackbeards” of a state could pirate software at will. Guess I won’t be writing any software for the state of North Carolina any time soon!
But what about content created by a state? Well, the state of Georgia makes various works available behind a paywall. That paywall might be run by a third party in exchange for a cut of the proceeds. So Public.Resource goes after anything where the edict of a government isn’t public domain. In other words, court decision, laws, and statutes should be free to all who wish to access them. The “government edicts doctrine” won in the end and so access to the laws of the nation continue to be free.
What about algorithms? That’s more patent territory when they are actually copyrightable, which is rare. Gottschalk v. Benson was denied a patent for a new way to convert binary-coded decimals to numerals while Diamond v Diehr saw an algorithm to run a rubber molding machine was patentable. And companies like Intel and Broadcom hold thousands of patents for microcode for chips.
What about the emergence of open source software and the laws surrounding social coding? We’ll get to the emergence of open source and the consequences in future episodes!
One final note, most have never heard of the names in early cases. Most have heard of the organizations listed in later cases. Settling issues in the courts has gotten really, really expensive. And it doesn’t always go the way we want. So these days, whether it’s Apple v Samsung or other tech giants, the law seems to be reserved for those who can pay for it. Sure, there’s the Erin Brockovich cases of the world. And lady justice is still blind. We can still represent ourselves, case and notes are free. But money can win cases by having attorneys with deep knowledge (which doesn’t come cheap). And these cases drag on for years and given the startup assembly line often halts with pending legal actions, not many can withstand the latency incurred. This isn’t a “big tech is evil” comment as much as “I see it and don’t know a better rubric but it’s still a thing” kinda’ comment.
Here’s something better that we’d love to have a listener take away from this episode. Technology is always changing. Laws usually lag behind technology change as (like us) they’re reactive to innovation. When those changes come, there is opportunity. Not only has the technological advancement gotten substantial enough to warrant lawmaker time, but the changes often create new gaps in markets that new entrants can leverage. Either leaders in markets adapt quickly or see those upstarts swoop in, having no technical debt and being able to pivot faster than those who previously might have enjoyed a first user advantage. What laws are out there being hashed out, just waiting to disrupt some part of the software market today?