Per Brinch Hansen: The Nucleus of a Multiprogramming System (1970) [pdf]
brinch-hansen.net
brinch-hansen.net
https://en.wikipedia.org/wiki/GEC_4000_series
Seems that Microsoft copied the idea in VerveOS with its Nucleus. Sure they improved on it but they patented the whole work. Good that I found the prior product to help invalidate those. The reason: I think the Nucleus concept is smart and especially making it the firmware. Like Wirth's system modules, it means everything above can be written in a safer, consistent language instead of a bunch of them. Plus, fewer abstraction gaps as everything uses same core model.
My concept is to look at each of those things GEC's Nucleus does, find the best work I can use in each, and integrate them into a modern one on a processor such as SAFE or CHERI. Then, put a language such as Ada, Go, or Ocaml on top of that. The result should be highly reliable.
I have a feeling nowadays most people would prefer being called the former over the later!
He insisted on us using his own book (i.e. not the "dragon" book) on compiler construction. His own material was utterly inferior, but he created a lock-in where we had to use his own little (inferior) operating system with a strange editor and a single, limited pascal compiler. We learned only the top-down recursive approach, ignoring bottom-up.
No fond memory. I had to study the Dragon book on my own just to get some perspective on compilers.
He conflated his business interests with his teaching, going as far as collecting stats on how many of his books had been purchased from the Uni bookstore, and reminding us at the beginning of each session that the book (his book) was obligatory and that photocopying was illegal (it was not) and that he would expel anyone caught studying using a copy of his book.