An Abridged History of Free And Open Source Software

    The History of Computing 11/24/2021


In the previous episodes, we looked at the rise of patents and software and their impact on the nascent computer industry. But a copyright is a right. And that right can be given to others in whole or in part. We have all benefited from software where the right to copy was waved and it’s shaped the computing industry as much, if not more, than proprietary software.

The term Free and Open Source Software (FOSS for short) is a blanket term to describe software that’s free and/or whose source code is distributed for varying degrees of tinkeration. It’s a movement and a choice. Programmers can commercialize our software. But we can also distribute it free of copy protections. And there are about as many licenses as there are opinions about what is unique, types of software, underlying components, etc. But given that many choose to commercialize their work products, how did a movement arise that specifically didn’t?

The early computers were custom-built to perform various tasks. Then computers and software were bought as a bundle and organizations could edit the source code. But as operating systems and languages evolved and businesses wanted their own custom logic, a cottage industry for software started to emerge. We see this in every industry - as an innovation becomes more mainstream, the expectations and needs of customers progress at an accelerated rate.

That evolution took about 20 years to happen following World War II and by 1969, the software industry had evolved to the point that IBM faced antitrust charges for bundling software with hardware. And after that, the world of software would never be the same.

The knock-on effect was that in the 1970s, Bell Labs pushed away from MULTICS and developed Unix, which AT&T then gave away as compiled code to researchers. And so proprietary software was a growing industry, which AT&T began charging for commercial licenses as the bushy hair and sideburns of the 70s were traded for the yuppy culture of the 80s. In the meantime, software had become copyrightable due to the findings of CONTU and the codifying of the Copyright Act of 1976.

Bill Gates sent his infamous “Open Letter to Hobbyists” in 1976 as well, defending the right to charge for software in an exploding hobbyist market. And then Apple v Franklin led to the ability to copyright compiled code in 1983. There was a growing divide between those who’d been accustomed to being able to copy software freely and edit source code and those who in an up-market sense just needed supported software that worked - and were willing to pay for it, seeing the benefits that automation was having on the capabilities to scale an organization.

And yet there were plenty who considered copyright software immoral. One of the best remembered is Richard Stallman, or RMS for short. Steven Levy described Stallman as “The Last of the True Hackers” in his epic book “Hackers: Heroes of the Computer Revolution.” In the book, he describes the MIT Stallman joined where there weren’t passwords and we didn’t yet pay for software and then goes through the emergence of the LISP language and the divide that formed between Richard Greenblatt, who wanted to keep The Hacker Ethic alive and those who wanted to commercialize LISP. The Hacker Ethic was born from the young MIT students who freely shared information and ideas with one another and help push forward computing in an era they thought was purer in a way, as though it hadn’t yet been commercialized.

The schism saw the death of the hacker culture and two projects came out of Stallman’s technical work: emacs, which is a text editor that is still included freely in most modern Unix variants and the GNU project. Here’s the thing, MIT was sitting on patents for things like core memory and thrived in part due to the commercialization or weaponization of the technology they were producing. The industry was maturing and since the days when kings granted patents, maturing technology would be commercialized using that system.

And so Stallman’s nostalgia gave us the GNU project, born from an idea that the industry moved faster in the days when information was freely shared and that knowledge was meant to be set free. For example, he wanted the source code for a printer driver so he could fix it and was told it was protected by an NDAQ and so couldn’t have it. A couple of years later he announced GNU, a recursive acronym for GNU’s Not Unix. The next year he built a compiler called GCC and the next year released the GNU Manifesto, launching the Free Software Foundation, often considered the charter of the free and open source software movement.

Over the next few years as he worked on GNU, he found emacs had a license, GCC had a license, and the rising tide of free software was all distributed with unique licenses. And so the GNU General Public License was born in 1989 - allowing organizations and individuals to copy, distribute, and modify software covered under the license but with a small change, that if someone modified the source, they had to release that with any binaries they distributed as well.

The University of California, Berkley had benefited from a lot of research grants over the years and many of their works could be put into the public domain. They had brought Unix in from Bell Labs in the 70s and Sun cofounder and Java author Bill Joy worked under professor Fabry, who brought Unix in. After working on a Pascal compiler that Unix coauthor Ken Thompson left for Berkeley, Joy and others started working on what would become BSD, not exactly a clone of Unix but with interchangeable parts. They bolted on the OSI model to get networking and through the 80s as Joy left for Sun and DEC got ahold of that source code there were variants and derivatives like FreeBSD, NetBSD, Darwin, and others. The licensing was pretty permissive and simple to understand:

Copyright (c) . All rights reserved.
Redistribution and use in source and binary forms are permitted provided that the above copyright notice and this paragraph are duplicated in all such forms and that any documentation, advertising materials, and other materials related to such distribution and use acknowledge that the software was developed by the . The name of the may not be used to endorse or promote products derived from this software without specific prior written permission.

By 1990 the Board of Regents at Berkley accepted a four clause BSD license that spawned a class of licenses. While it’s matured into other formats like a 0 clause license it’s one of my favorites as it is truest to the FOSS cause.

And the 90s gave us the Apache License, from the Apache Group, loosely based on the BSD License and then in 2004 leaning away from that with the release of the Apache License 2 that was more compatible with the GPL license. Given the modding nature of Apache they didn’t require derivative works to also be open sourced but did require leaving the license in place for unmodified parts of the original work.

GNU never really caught on as an OS in the mainstream, although a collection of tools did. The main reason the OS didn’t go far is probably because Linus Torvalds started releasing prototypes of his Linux operating system in 1991. Torvalds used The GNU General Public License v2, or GPLv2 to license his kernel, having been inspired by a talk given by Stallman. GPL 2 had been released in 1991 and something else was happening as we turned into the 1990s: the Internet. Suddenly the software projects being worked on weren’t just distributed on paper tape or floppy disks; they could be downloaded. The rise of Linux and Apache coincided and so many a web server and site ran that LAMP stack with MySQL and PHP added in there. All open source in varying flavors of what open source was at the time.

And collaboration in the industry was at an all-time high. We got the rise of teams of developers who would edit and contribute to projects. One of these was a tool for another aspect of the Internet, email. It was called popclient, Here Eric S Raymond, or ESR for short, picked it up and renamed it to fetchmail, releasing it as an open source project.

Raymond presented on his work at the Linux Congress in 1997, expanded that work into an essay and then the essay into “The Cathedral and the Bazaar” where bazaar is meant to be like an open market. That inspired many to open source their own works, including the Netscape team, which resulted in Mozilla and so Firefox - and another book called “Freeing the Source: The Story of Mozilla” from O’Reilly.

By then, Tim O’Reilly was a huge proponent of this free or source code available type of software as it was known. And companies like VA Linux were growing fast. And many wanted to congeal around some common themes. So in 1998, Christine Peterson came up with the term “open source” in a meeting with Raymond, Todd Anderson, Larry Augustin, Sam Ockman, and Jon “Maddog” Hall, author of the first book I read on Linux. Free software it may or may not be but open source as a term quickly proliferated throughout the lands.

By 1998 there was this funny little company called Tivo that was doing a public beta of a little box with a Linux kernel running on it that bootstrapped a pretty GUI to record TV shows on a hard drive on the box and play them back. You remember when we had to wait for a TV show, right? Or back when some super-fancy VCRs could record a show at a specific time to VHS (but mostly failed for one reason or another)? Well, Tivo meant to fix that. We did an episode on them a couple of years ago but we skipped the term Tivoization and the impact they had on GPL.

As the 90s came to a close, VA Linux and Red Hat went through great IPOs, bringing about an era where open source could mean big business. And true to the cause, they shared enough stock with Linus Torvalds to make him a millionaire as well. And IBM pumped a billion dollars into open source, with Sun moving to open source Now, what really happened there might be that by then Microsoft had become too big for anyone to effectively compete with and so they all tried to pivot around to find a niche, but it still benefited the world and open source in general.

By Y2K there was a rapidly growing number of vendors out there putting Linux kernels onto embedded devices. TiVo happened to be one of the most visible. Some in the Linux community felt like they were being taken advantage of because suddenly you had a vendor making changes to the kernel but their changes only worked on their hardware and they blocked users from modifying the software. So The Free Software Foundation updated GPL, bundling in some other minor changes and we got the GNU General Public License (Version 3) in 2006.

There was a lot more in GPL 3, given that so many organizations were involved in open source software by then. Here, the full license text and original copyright notice had to be included along with a statement of significant changes and making source code available with binaries. And commercial Unix variants struggled with SGI going bankrupt in 2006 and use of AIX and HP-UX

Many of these open source projects flourished because of version control systems and the web. SourceForge was created by VA Software in 1999 and is a free service that can be used to host open source projects. Concurrent Versions System, or CVS had been written by Dick Grune back in 1986 and quickly became a popular way to have multiple developers work on projects, merging diffs of code repositories. That gave way to git in the hearts of many a programmer after Linus Torvalds wrote a new versioning system called git in 2005. GitHub came along in 2008 and was bought by Microsoft in 2018 for 2018.

Seeing a need for people to ask questions about coding, Stack Overflow was created by Jeff Atwood and Joel Spolsky in 2008. Now, we could trade projects on one of the versioning tools, get help with projects or find smaller snippets of sample code on Stack Overflow, or even Google random things (and often find answers on Stack Overflow). And so social coding became a large part of many a programmers day. As did dependency management, given how many tools are used to compile a modern web app or app. I often wonder how much of the code in many of our favorite tools is actually original.

Another thought is that in an industry dominated by white males, it’s no surprise that we often gloss over previous contributions. It was actually Grace Hopper’s A-2 compiler that was the first software that was released freely with source for all the world to adapt. Sure, you needed a UNIVAC to run it, and so it might fall into the mainframe era and with the emergence of minicomputers we got Digital Equipment’s DECUS for sharing software, leading in part to the PDP-inspired need for source that Stallman was so adamant about. General Motors developed SHARE Operating System for the IBM 701 and made it available through the IBM user group called SHARE. The ARPAnet was free if you could get to it. TeX from Donald Knuth was free. The BASIC distribution from Dartmouth was academic and yet Microsoft sold it for up to $100,000 a license (see Commodore ). So it’s no surprise that people avoided paying upstarts like Microsoft for their software or that it took until the late 70s to get copyright legislation and common law.

But Hopper’s contributions were kinda’ like open source v1, the work from RMS to Linux was kinda’ like open source v2, and once the term was coined and we got the rise of a name and more social coding platforms from SourceForge to git, we moved into a third version of the FOSS movement. Today, some tools are free, some are open source, some are free as in beer (as you find in many a gist), some are proprietary. All are valid.

Today there are also about as many licenses as there are programmers putting software out there. And here’s the thing, they’re all valid. You see, every creator has the right to restrict the ability to copy their software. After all, it’s their intellectual property. Anyone who chooses to charge for their software is well within their rights. Anyone choosing to eschew commercialization also has that right. And every derivative in between. I wouldn’t judge anyone based on any model those choose. Just as those who distribute proprietary software shouldn’t be judged for retaining their rights to do so.

Why not just post things we want to make free? Patents, copyrights, and trademarks are all a part of intellectual property - but as developers of tools we also need to limit our liability as we’re probably not out there buying large errors and omissions insurance policies for every script or project we make freely available. Also, we might want to limit the abuse of our marks. For example, Linus Torvalds monitors the use of the Linux mark through the Linux Mark Institute. Apparently some William Dell Croce Jr tried to register the Linux trademark in 1995 and Torvalds had to sue to get it back. He provides use of the mark using a free and perpetual global sublicense. Given that his wife won the Finnish karate championship six times I wouldn’t be messing with his trademarks.

Thank you to all the creators out there. Thank you for your contributions. And thank you for tuning in to this episode of the History of Computing Podcast. Have a great day.

(OldComputerPods) ©Sean Haas, 2020