Program design in the Unix environment (1984) [pdf]

Somewhat off-topic, but now that Uriel is dead, is there any central figure tasking themselves with the future upkeep of cat-v?

"Unix" as we know it is a merger of two traditions - the spartan purism of Research Unix and the rich user experience of PDP-10 OSes (Tenex, ITS, WAITS, etc.). When DEC killed the PDP-10 a lot of the PDP-10 culture jumped to BSD, bringing a few big programs (TeX and Emacs) and a few popular features (command completion) which the Bell folks either wouldn't have thought to add or didn't want.

This is one of many times when Rob Pike has complained about non-Research Unix being "impure" (and also VT100s suck and the Blit is the pinnacle of GUI design). He's right to a point, but an OS that enforced purity wouldn't have had nearly the success Unix has had. Most people don't want pure, they want useful, and purity can help usefulness or get in the way.

Wow, they mention 9600 baud being common in 1984, but I remember having a brand new 9600 baud modem around 1991, I guess we were late to the game :)

It was interesting to see a reference to James Gosling's "UNIX Emacs" paper (reference (3) at the end). I keep getting reminded every once in a while that he did a lot of things before Java.