The foundations of a provably secure operating system (PSOS) (1979) [pdf]

We had a nearly unhackable computer plan in 1979, and commenters are furious we blew it

TLDR: A 1979 paper described a computer system built to be secure from the ground up, with rules so strict they could be formally proven. Commenters are split between awe and rage, arguing we ignored a safer path and spent decades patching weaker systems instead.

A dusty 1979 paper about a "provably secure" operating system somehow turned into a full-blown "we chose the bad timeline" moment in the comments. The paper itself is pretty wild: researchers at SRI sketched a computer system where every file, program, and process could only be touched with special unchangeable "keys," and the whole thing was designed so its safety rules could actually be proven, not just hoped for. In plain English, this was an early attempt to build a computer that was secure by design, not patched together after the fact.

But the real fireworks came from readers who looked at this 1979 blueprint and basically screamed, "Wait, we had this all along?" One of the loudest reactions argued that the industry threw away a path to true built-in security for the sake of convenience, calling it a lost chance at hardware-level zero-trust. That comment caused its own mini-drama when another user popped in to ask why it had even been flagged, adding a side plot of comment moderation mystery to the nostalgia spiral.

Then came the modern-day doomposting: one commenter said this kind of design may have seemed overkill in the 1970s, but once the internet let people download random software from anywhere, it became the obvious model—and yet we kept piling hacks on top of old systems instead. And because no tech thread is complete without an "AI will rebuild civilization" cameo, another user suggested we should start over, let AI agents write everything, and then make them prove it works. Sensible? Maybe. Chaotic? Absolutely.

Key Points

  • The 1979 PSOS paper describes an operating system designed using the SRI Hierarchical Development Methodology (HDM) and formal verification techniques.
  • HDM in PSOS uses formal requirements, formal module specifications, and formalized module interconnections to support proofs of correctness.
  • PSOS was specified in the SPECIAL language as about 20 hierarchically organized modules covering capabilities, virtual memory segments, directories, user processes, and user-defined abstract objects.
  • The system uses capabilities as its uniform protection mechanism, requiring presentation of a valid capability to access any object.
  • Each PSOS capability contains an immutable unique identifier and immutable access rights, and capabilities are described as non-forgeable and non-alterable.

Hottest takes

"we had the blueprint for true hardware-level Zero-Trust in 1979, and we abandoned it for deployment convenience" — Miagg
"the only architecture suitable for the internet age" — usrbinenv
"Rebuild everything from scratch, with AI agents. Then make them prove what they wrote." — GistNoesis
Made with <3 by @siedrix and @shesho from CDMX. Powered by Forge&Hive.