back
3 comments
Minix is not dead, it has become bloated and thus lost relevance.

The original purpose of Minix was as an operating system for teaching purposes. The NetBSD userland, etc has lead to a massive amount of bloat. Somewhere along the line, people have forgotten that Minix was the launchpad for Linux. So if you want a full-featured OS, then Linux or NetBSD are more appropriate.

Minix as the core of Intel's management processor is a distraction. Most of work by Intel has been kept in-house. For what it is worth, Intel's work is probably of minimal value to Minix hackers.

For learning purposes, the original Minix release and the Tanenbaum & Woodhull book remains a great resource.

From an architectural point of view, Minix has a microkernel at the core and is more like Mach, or seL4. Device drivers are in userland, and thus are more robust. Tanenbaum's funding to create a high reliability European OS became a distraction that lead to nowhere.

IMHO moving away from the Amsterdam Compiler to GCC introduced a lot of incidental complexity. I would have preferred adopting the Plan9 C compilers and thus gaining easy cross-compilation capabilities.

Recently discussed: https://news.ycombinator.com/item?id=36064961
It's pining for the fjords.