Historically, the operating system was largely used as a research platform for writing papers and theses [2]. Unfortunately, research code isn't production-grade code, so despite a lot of effort done to mainline some of it, not all of it was merged. Also, while there was a steady supply of new students to work on MINIX3, there was also a steady turnover because students who finish their studies generally move on to new things.
With the looming retirement of Andrew Tanenbaum in 2014, the project pivoted focus to embedded systems, due to the technical strengths and weaknesses of the system at that time. The ARM port was done around that time as part of it to leverage small board computers. Eventually, the funding dried up and the project slowed to a standstill, largely I believe due to an unsustainable maintenance burden caused by accumulated tech debt over decades that the remaining development manpower (all contributors working on their free time, including myself) couldn't address.
MINIX3 still has some unique cool stuff to this day like seamless live updates of system components, but for me it's too shackled by its past to have a future. It'd be a lot of work to bring it up to modern standards and personally by the time I'd be done with that I think ought to be done, it wouldn't look like MINIX anymore.
If you were teaching an OS class in the late 1980s/early 1990s and wanted to have students work on and compile the source code of a UNIX-like system, using desktop PCs of the era, Minix was pretty much it wasn't it?
At that time, MINIX had a proprietary license (it wasn't open-source until 2000) and its stated purpose was being a teaching tool for education, not being a production operating system. I can't find the quote itself, but I recall Tanenbaum once said something along the lines of "Turns out I do want to turn MINIX3 into a product, it just took me thirty years to realize it!" (please don't quote me on that unless I manage to dig up the actual quote).
> If you were teaching an OS class in the late 1980s/early 1990s and wanted to have students work on and compile the source code of a UNIX-like system, using desktop PCs of the era, Minix was pretty much it wasn't it?
Nowadays, xv6 fills that niche. Recent versions of MINIX3 arguably outgrew that with the switch over to the NetBSD userland.
Thanks, I didn't remember that but now that you say it that sounds right. Still, I think it was one of the few ways students at the time could inspect the code of a complete OS, modify it, recompile it, and run their changes, all on the PC that many of them had at home or in classrooms.
However, my point is not about copyright,etc. I'm simply pointing out, as Linus has clearly stated in the early days, that he booted a system with Minix and used it to start up the development of what he ended up calling Linux. Over the 30+ years since then, I would suspect that any carryover from Minix is long gone. Even support for Minix filesystem appears (to me) to be deprecated.
The considerable bloat detracted from the principle of using Minix to teach OS principles.
The real question is whether the project is deader than HURD.
[1] https://web.archive.org/web/20240920163544/https://www.cs.vu.nl/~ast/intel/
That's hardly relevant though. Either a project is usable and people use it, or a project is unusable and people don't care if it even exists.
Is there actually an official Intel document which confirms this? So far I only saw claims by third parties. Do we know which version of Minix was used, and which parts of it?
The latter also still has people (outside IBM) working on it. It appears to have changed hands a few times.
https://www.arcanoae.com/arcaos/