Commercial open source software has always been a head scratcher for many: how exactly can you publish your entire code base online, make it free for people to download, but also create a viable business around it? Red Hat's IPO in '99 was a great result, but it was more of a one-of success than a playbook that other makers could follow. In the past decade open source has boomed in popularity, but the path to commercialization has gotten even more complicated, as the shift to the cloud put AWS and other providers in a position to easily host and redistribute open source project without rewarding the original creators. In this post we will dive into the history of open source software licensing, its evolution over the last 40 years, and what the future holds.
Before the 70s, software was not considered copyrightable, and therefore had no licenses attached to it; early commercial products like IBM's mainframe software were distributed with source code included. The landscape started changing with the "US Commission on New Technological Uses of Copyrighted Works (CONTU)" of 1974, which stated that "computer programs, to the extent that they embody an author's original creation, are proper subject matter of copyright". In 1976, a young Bill Gates authored "Open Letter to Hobbyists" and published it on the New York Times; in the essay, Gates denounced people who obtained copies of BASIC without paying the license fee: "Who can afford to do professional work for nothing?". These copies were bootleg versions of the BASIC paper tape. One of the original tapes was stolen at a Homebrew Computer Club meetup in June 1975, and copies of it were redistributed at the following meeting.
The final nail in the coffin of public-domain software came with the verdict of *[Apple vs Franklin](https://en.wikipedia.org/wiki/Apple_Computer,_Inc._v._Franklin_Computer_Corp.)* in 1983, which clarified that both the source code (the human-readable code) and the binary code (the machine readable code) of a program are covered by copyright. After Apple vs Franklin, the software landscape quickly started moving towards widespread licensing and binary-only distribution of software. IBM reverted its original policy and announced it'd be no longer distributing the source code along with software purchases.
"Free software" was now vulnerable to larger companies leveraging their source code and redistributing it as a commercial product. Apple itself was a clear representation of the divide between hobbyists, like Steve Wozniak and the Homebrew Computer Club, and more commercially-focused minds, like Steve Jobs and Bill Gates.
In the 1970s Richard Stallman, a Harvard graduate, spent time at MIT's AI laboratory where he built a reputation as a strong proponent of "free software". Disturbed by the development in the software licensing world, he founded the GNU Project in 1983, with the aim of representing and supporting the free software movement. The organization is responsible for the creation of the GNU General Public License (GPL) in 1989, which would then be adopted by leading projects of the time like Linux and MySQL.
The GPL leveraged a new concept called "copyleft". "Copyleft" allows software creators to grant users access to the source code of their applications and lets them create derivative work on top of it, but it requires that work to also adopt the same license. This makes it impossible for companies to take projects under "copyleft", package them and redistribute them as closed-source commercial products. The GPL was a pivotal moment in the history of open source software; to this day, the license is still leveraged by some of the most popular open source projects like Red Hat, MySQL, Git, and Ubuntu.
In 1997 Eric Raymond, a long-time Linux contributor, published "The Cathedral and the Bazaar", in which he talked about his experience in free software (the "Bazaar" model), and what his advantages are over the status quo of packaged software the industry offered (the "Cathedral" model). In summary, Raymond argues that the wisdom of the crowd offered by the Bazaar model generates better results than a small group of developers could. "Given enough eyeballs, all bugs are shallow." Netscape's CEO would mention Raymond's work as one of his inspirations when deciding to open source the browser (more on that later).
In that same year Christine Peterson, a fellow MIT alum, coined the term "open source" during a group meeting with Raymond and other "free software" leaders at the Foresight Institute. We should all consider ourselves lucky that "cooperatively developed" didn't win that fight! Raymond updated his essay accordingly on February 9th 1998 ("Changed free software to open source'') and at the "Freeware Summit" of 1998, a vote was taken to adopt "Open source" as the standard term. The event was then renamed to "Open Source Summit", which still runs to the this day, supported by the Linux Foundation.
In March 1998, Netscape open sourced their Communicator 5.0 source code on their newly registered Mozilla.org domain. This quote from Bob Lisbonne, SVP of client products at Netscape at the time, touches on one of the most important aspects of open source: the possibility of tapping into an unlimited pool of talent.
"It's no longer Netscape alone, pushing the client software forward, but now it's really the whole Net. For Netscape, this gives us a way to engage the creative, innovative abilities of literally orders of magnitude more people than we could ever--really any commercial software company could ever--afford to just put on their payroll."
The software was released under the Netscape Public License, which wasn't recognized as a true open source one, and received criticism by Richard Stallman. Mozilla eventually spun out of Netscape, and launched the Firefox browser in 2002 under the Mozilla Public License. In that same time period, other open source projects were being launched both in enterprise contexts, like Sun Microsystems's OpenOffice.org and Rapid7's Metasploit, but also in the gaming space, like Linden Lab's Second Life. In other cases, the open source project was preceding the company: Matt Mullenweg created WordPress in 2003, but only left CNET to start Automattic in 2006, and raised $1.1M in 2007.
In the early 2000s, licensing was still pretty straight forward, with most projects living under GPL or BSD licenses. The latter (not to be confused with Linux's BSD distribution) is one of the most permissive licenses, and doesn't require the disclosure of source code of derivative works. For example, I could take an open source BSD-licensed project, modify it, and redistribute it as commercial closed source product without issues.
GPL licenses on the other hand require you to disclose your source code if it leverages other GPL-licensed pieces of code. While great in theory, this is hard to enforce; if a closed source application uses GPL source code but I have no visibility into the code itself, how can I confirm it's in violation? Free Software Foundation (FSF) vs Cisco in 2008's was one of the first GPL lawsuits that were against a large enterprise. In 2016 Automattic accused Wix of using GPL-licensed code in one of their projects, but quickly resolved the issue out of court.
As cloud computing became more popular in the 2000s, there was a major shift in the open source world. In the early days, a big part of the OSS community was individual developers participating in the development of these projects as a hobby; as time went on, open source became a larger focus for enterprises. This usually happens in two ways:
This shift caused GPL to lose in popularity, as it'd require companies to open source their core products if they were to use their own open source work. Companies started moving to an Apache or MIT license. Nowadays, those two licenses account for more than 50% of all licenses. This also allowed the rise of startups that can help companies manage open source compliance, such as FOSSA and Snyk.