A Short History of Open Source

To most people “open source” feels like an invention of the internet age. In fact it is the other way round: in the early years of software, sharing was normal and closedness came later. This post walks through the turning points of that long story, because today’s debate about “open weights” in artificial intelligence only makes sense against this background.

When sharing was normal

How was software shared in the early years of computing?

In the 1950s and 60s, software was an add-on that came with the hardware. IBM users founded a group called SHARE in 1955 and shared their software and fixes with one another within it. In academic laboratories, code circulated from hand to hand like papers. Publishing a method meant sharing its program as well.

In those years, sharing source code was not a separate ideology; it was simply part of the work.

The age of ownership

When did software become a product for sale?

In the 1970s, software became a product separate from hardware and could be sold. The document seen as the symbol of this period is the open letter Bill Gates wrote to computer hobbyists in 1976. In it Gates argued that, because software was being copied without permission, developers were not paid for their work and this prevented good software from being written. That debate continues today. But in those years the direction of the industry was clear: code coming out of university laboratories was increasingly withdrawn behind commercial licences.

GNU and copyleft

How did the free software movement begin?

The first systematic answer to this trend came in 1983. Richard Stallman announced the GNU project, which aimed at a complete, free, Unix-like operating system. In 1985 he founded the Free Software Foundation (FSF). In the same years a legally inventive idea appeared: copyleft. First tried in the GNU Emacs licence, the idea was generalised in 1989 with the GPL. The aim was to use copyright in reverse. Anyone who received the code had to grant the same freedoms to others when redistributing it.

According to the FSF’s definition, “free” should be understood as in “free speech”, not as in “free beer”. In other words, the point is not the price but the user’s freedom.

Linux: from hobby to ecosystem

How did Linux become so widespread?

In 1991 Linus Torvalds, then a student at the University of Helsinki, made a modest announcement on Usenet. He said he was writing a (free) operating system “just a hobby” and that it “won’t be big and professional like gnu”. That sentence may count as one of the least accurate predictions in the history of technology.

When the Linux kernel met the GNU tools and settled into the legal framework of the GPL, a self-sustaining ecosystem emerged. In 1998 Netscape announced that it would open the source code of its browser. After that the term “open source” was coined and the Open Source Initiative (OSI) was founded. IBM’s announcement at the end of 2000 of an investment of about one billion dollars in Linux for 2001 showed that the business world was taking the movement seriously too.

GitHub and social coding

How did GitHub change open source?

Founded in 2008, GitHub turned contributing into a social interaction. Concepts such as “fork” and “pull request” became the language of shared production. Version control stopped being an engineering detail and became the way developers communicate with one another. Open source was no longer only a choice of licence; it was a culture of work.

Open weights: a new chapter of an old debate

Should the weights of AI models be shared?

One of the big questions of the 2020s is whether the weights of large language models should be shared. The split between advocates of “open weights” and those who consider closed models safer recalls the software ownership debate of the 1970s to a surprising degree: labour or sharing? What kind of sharing should be required, for whom, and to what extent?

I think history offers a clue here. When we look at the history of science, from correspondence to peer-reviewed journals, we see a tradition in which knowledge accumulates as it is shared. The last forty years of software point in a similar direction.

Closing

What do we learn from the history of open source?

The story of open source is not a victory march along a straight line. It is a long debate that has swung back and forth. But one tendency is clear: shared knowledge accumulates. Perhaps in a future post we will take a closer look at the “open source” traditions of the history of science itself, from letters to archives.

References

1. Free Software Foundation. (n.d.). What is free software? The free software definition. GNU Project. https://www.gnu.org/philosophy/free-sw.html
2. Kelty, C. M. (2008). Two bits: The cultural significance of free software. Duke University Press.
3. Open Source Initiative. (n.d.). History of the OSI. https://opensource.org/history
4. Stallman, R. (1983, September 27). Initial announcement [Usenet post, net.unix-wizards]. GNU Project. https://www.gnu.org/gnu/initial-announcement.html
5. Weber, S. (2004). The success of open source. Harvard University Press. https://doi.org/10.4159/9780674044999