On January 29, 1992, Andrew Tanenbaum started a thread in the comp.os.minix newsgroup with the blunt subject line “LINUX is obsolete.” His opening was almost casual: “I was in the U.S. for a couple of weeks, so I haven’t commented much on LINUX… but for what it is worth, I have a couple of comments now.” The full exchange is reproduced as an appendix in the O’Reilly book “Open Sources.”
Tanenbaum’s argument was about architecture. He held that a microkernel, like his own MINIX, was the better design because it kept the kernel small and portable, and he criticized Linux for being a monolithic kernel tightly bound to the Intel 386. In his view, tying a system so closely to one processor family made it obsolete the moment the hardware changed.
Torvalds answered within a day. He conceded the theory, writing that “from a theoretical (and aesthetical) standpoint linux looses,” but argued that a monolithic kernel was simpler and faster to build for the machine he actually had, and he pointed out design limits in MINIX itself. Other well-known programmers, including David Miller and others on the newsgroup, joined in.
The thread became a touchstone for the culture of systems programming, the classic statement of the microkernel-versus-monolithic argument. Tanenbaum later took pains to say the fight was technical, not personal: “First, I REALLY am not angry with Linus,” he wrote, adding “He’s not angry with me either.” Linux, the supposedly obsolete design, went on to dominate.