To answer the question of the future of open source, I think it's important to consider why open source has existed for so long. Is it a matter of cost to develop closed-source systems? Or is it about the quality and efficiency enabled by open source development?
My experience is that the 0% interest on credit time for large companies enabled open source to thrive: open source maintainers were flush with cash and time. Large companies with ample cash let their staff do anything to keep them happy (and away from their competitors!), including allocated time for OSS contributions.
Now that 0% interest has ended 'regular' people like me are not flush with cash. Any time I have I need to spend on activities that will bring in money. Why would I waste that on open source?
(Another aspect is the McKinsey-ification of the work place in the last ~10 years or so. Managers are making decisions in tech now, not tech people. all my life I was told that OSS contributions will look great on my CV. So far nobody who has made a decision to hire me has had the background, interest, or knowledge to judge, or even care, for my OSS contributions.)
It hasn't existed for so long! Jeez, kids these days. As a movement it's really only about 30 years old, and that in two separate phases:
1. The GPL/GNU period. The software is mostly called "free software-as-in-freedom". The community is small, volunteer driven, highly ideological and focused primarily on cloning UNIX for commodity PC hardware. Most developers aren't using this stuff at all and work with entirely commercial toolchains.
2. The Apache 2 period. The software is mostly called "open source". The community is huge, frequently driven by corporate donations, not particularly ideological anymore and focused on developer tools / libraries rather than operating systems and desktop apps. Developers now regularly incorporate open source libraries into their commercial programs.
The transition was slow but I'd pick 2008-2010 as the crossing over point. Before that time if someone said they wrote open source software you'd make a weak assumption that they worked somehow on Linux related projects in their evenings/weekends. After that time you'd probably assume they were writing some library and had even odds it was their job.
It's important to remember that 2008+ is ZIRP territory. Money was "free" for investors, so investment piled into a lot of stuff that often wasn't revenue generating because there was nowhere else for it to go. This era was also shaped by a historical aberration - a handful of hyper successful advertising companies whose founders were able to remain in control whilst still selling much of their stock thanks to dual voting classes. Modern developer's understanding of the open source ecosystem, along with their expectation that powerful tools are all free, is very much shaped by the combination of Fed policy and a handful of super rich patron companies that weren't under any pressure to return capital to shareholders.
Thing is none of these factors (ideology, free money, stock with dual voting classes) is historically normal or arguably sustainable over the long run, whereas ordinary capitalism is. So we might well see a reversion to the mean where things like compilers, libraries, operating systems etc become commercial again as relatively high interest rates pull funds out of tech and the supply of Stallmanists devoted to the cause of desktop GNU/Linux continues to dwindle. The recent death of a man who reverse engineered a lot of networking hardware for Linux is an example of this - how many 25 year old hackers want to do that sort of thing any more? They're all writing JS frameworks that only work on AWS these days.
If that does happen though it'll be quite slow. I think the industry would need a kind of Steam for libraries to emerge first, and it's pretty unclear what the next equilibrium phase looks like.
30 years is the majority of the existence of personal computers have been available. It’s a long time
> how many 25 year old hackers want to do that sort of thing?
A lot! More than 30 years ago, that’s for sure. The sheer number of all programmers have increased so much since then.
A few years ago, when I was 25, I was getting into hardware hacking and I found 2 books on the subject.
It’s just hard to find learning resources on reverse engineering hardware, since that isn’t the entry point for programmers anymore.
My hope is that some greybeards will write some resources on how they made harfbuzz or eMacs or whatever.
I’d pay for that knowledge. I’d gladly pay for technical biographies of open source projects.
I was being pretty generous with my timespans. 30 years ago was 1994. Personal computers had existed for quite a long time by then. The Apple Mac launched in 1984.
Open source did exist in 1994: projects like Linux and Python were started around 1991 but nearly nobody knew about them or used them. The average person or developer in 1994 had zero encounters with open source software. Even by 2000 this was still the case: the average developer was working with Visual Basic or Delphi or Visual Studio (all proprietary) using the Windows API or VBX/OCX controls as libraries (proprietary), connecting to Oracle, SQL Server or Access for data (proprietary) and if they were bold, rendering web UIs from IIS or Netscape's servers (proprietary) to Internet Explorer or Netscape Communicator (all proprietary). If they were cutting edge like Google they might be using Linux as a server kernel, but that was rare and a source of competitive advantage. Google wrote all of its own internal libraries starting from the STL upwards partly because there just weren't that many to adopt if you worked on Linux.
> It’s just hard to find learning resources on reverse engineering hardware, since that isn’t the entry point for programmers anymore.
Well, two books is a lot. Back in the day there were none :) I think this shows the issue, right? Yes there are more developers overall, but there are also more opportunities and things to do. Why would you spend hours slaving over a buggy wifi driver when you could try your hand at writing a mobile app instead, which might make you rich? Back in the 90s this was far less of an option for most developers. The original motivation was a desire to use something other than Windows on PC class hardware, but that desire has been satiated by Apple for a long time and desktop Linux remains obscure.
I've experienced all of this. My first open source project was on Windows, back when open source was novel and new. Then I worked on Linux for a while - I had code in Wine and GNOME and a few other things. Then I wrote open source libraries and took part in Bitcoin. These days I do open source work for pay and also sell a proprietary developer tool. So, seen it from every angle. My gut feeling is that we're going to see a resurgence of proprietary platforms and libraries in the coming years.
> Why would you spend hours slaving over a buggy wifi driver when you could try your hand at writing a mobile app instead, which might make you rich?
People do get paid to write Linux WiFi drivers. Especially for embedded devices. But honestly, some people just like that stuff. I’ve watched a twitch stream where a hobbyist reverse engineered the wireless protocol that the valve knuckles controller used.
He spent hours sifting through bits in wireshark.
Some people see a problem and bash their heads against it until it’s fixed. Torvalds is one of those people and I’m sure many open source contributors are too.
Kind of an aside, but what was really cool was that the creator of that protocol was in chat dropping hints.
IMO the big failing of the open source community is selling out too much and not attracting as many enthusiasts as career programmers (not that there isn’t overlap)
Alan Perlis has a quote that goes like “I hope we keep the fun in computing” and I think we failed at that.
My outlook isn’t as bleak. I think Linux is here to stay. I don’t really see a proprietary OS that is as robust and runs on as many different platforms taking Linux’s thunder.
Mobile and VR has android.
Even on desktop, the seam deck is pushing it forward, though it will forever probably remain niche on desktop.
Sure, but the only reason Linux wifi drivers are open source is that Linux was born at the start of the GNU movement and Linus picked the GPL. If he had picked a BSD license, or LGPL for drivers, then people would still get paid to write them but they wouldn't be open source. What you see nowadays is that the GPL is long since abandoned, I don't remember the last time I encountered an open source project that was GPL outside of the few big ones that survive from the 90s.
With respect to maintainers getting old, I guess your example is kind of on point. It used to be driven by ideology or the practical desire for a home UNIX. That motivated people to spend the long evenings alone working it out. Now it's driven by building a Twitch audience, and the hobbyist was even getting help!
I guess with respect to stealing Linux's thunder, I feel like that happened 20 years ago already. When MacOS X came out there was a long stream of people 'defecting' to it from the Linux community. It separated those who just wanted UNIX from those who had the ideology and a lot of the energy dissipated from Linux at that time.
Open source has existed for as long as the Internet has.
And both were called by different names previously.
It is not wrong to say that open source grew up in parallel with the Internet. I see no reason to worry that they will not continue to evolve together.
There is another major aspect not discussed here and that is the proliferation of OSS libraries developed in academia which become widely used.
From my experience it’s about cost and efficiency. By cost I mean my cost to them for my time coding. Customers don’t factor in the hidden costs of open source.
I preferred developing minimal-dependency software but my customers demand fast good-enough results. The only way to deliver that is to glue together open source dependencies.