Live data from Hacker News

The Joys of Unix – NSA Cryptolog (1978)

archive.org

11–17 of 17 posts

Re: The Joys of Unix – NSA Cryptolog (1978)

#11
post #10

Does anyone happen to know why the shell got named a "shell"?

Just speculation, but perhaps it's because it "wraps" the kernel. ie. the kernel's shell...

That was always my understanding. In older mainframe operating systems, the command interpreter was more integral to the OS. In Unix, it's just a process. Its function is to run programs for the user, so it is in some sense a shell around the OS seen by the user.

Re: The Joys of Unix – NSA Cryptolog (1978)

#12
post #4

"I became, in fact, less and less of a programmer at all, and more and more simply a procedure-writer who tacked together canned routines or previously debugged POGOL steps to do dull things in a dull way". This was when the author had to work with layers of accidental complexity of the IBM 360s, before discovering Unix. It gives hope to know that such complexity existed even in 1978, but that it was simply bad engin…

On a similar note, there is the Niklaus Wirth rant about complexity as well on his ACM Turing Award speech.

http://dl.acm.org/ft_gateway.cfm?id=1283941&type=pdf

To the point his Oberon-07 redesign had as goal a minimalist GC systems programming language.[0]

http://www.inf.ethz.ch/personal/wirth/Oberon/Oberon07.pdf

http://www.inf.ethz.ch/personal/wirth/Oberon/Oberon07.Report...

[0] Too minimalist actually, as many of us might say.

Re: The Joys of Unix – NSA Cryptolog (1978)

#14
post #4

"I became, in fact, less and less of a programmer at all, and more and more simply a procedure-writer who tacked together canned routines or previously debugged POGOL steps to do dull things in a dull way". This was when the author had to work with layers of accidental complexity of the IBM 360s, before discovering Unix. It gives hope to know that such complexity existed even in 1978, but that it was simply bad engin…

This got me thinking: I wonder if places like St. John's College[1] might eventually take up teaching directly from Turing and von Neumann. Maybe one day those sorts of great-books focused liberal arts colleges will be the only places left teaching the low-level stuff and the basic principles of computing, outside a handful of hyper-focused engineering programs designed to meet the limited need for some people in the industry who understand it well.

[1] https://en.wikipedia.org/wiki/St._John%27s_College_(Annapoli...

(I can't speak to the rigor of that program from direct experience, but look at the reading list—it's not all the fluffy philosophy and literary criticism that many imagine when they think of liberal arts programs. They cover the major Greek mathematicians in the first two years, then it's off to enlightenment-era mathematics—John von Neumann or Turing, and maybe even something like TAoCP, wouldn't be out of place)

Re: The Joys of Unix – NSA Cryptolog (1978)

#15
post #10

Does anyone happen to know why the shell got named a "shell"?

Just speculation, but perhaps it's because it "wraps" the kernel. ie. the kernel's shell...

Yup, I've always thought it provides sort of an "airlock" generic land between userspace and kernel-land, where people can do things like send signals to processes and also do stuff like dump the contents of a file for easy viewing.

Re: The Joys of Unix – NSA Cryptolog (1978)

#16
post #10

Earlier quoted context omitted.

Just speculation, but perhaps it's because it "wraps" the kernel. ie. the kernel's shell...

That was always my understanding. In older mainframe operating systems, the command interpreter was more integral to the OS. In Unix, it's just a process. Its function is to run programs for the user, so it is in some sense a shell around the OS seen by the user.

That's the way I see it and to extend the analogy a bit further Unix, et el. are like a hermit crab that can trade one shell for another if there is a good reason to do so.

Re: The Joys of Unix – NSA Cryptolog (1978)

#17
post #14
post #4

"I became, in fact, less and less of a programmer at all, and more and more simply a procedure-writer who tacked together canned routines or previously debugged POGOL steps to do dull things in a dull way". This was when the author had to work with layers of accidental complexity of the IBM 360s, before discovering Unix. It gives hope to know that such complexity existed even in 1978, but that it was simply bad engin…

This got me thinking: I wonder if places like St. John's College[1] might eventually take up teaching directly from Turing and von Neumann. Maybe one day those sorts of great-books focused liberal arts colleges will be the only places left teaching the low-level stuff and the basic principles of computing, outside a handful of hyper-focused engineering programs designed to meet the limited need for some people in the…

The on St. John's grad I know well used to have a copy of a book including "On Computable Numbers" on his mantelpiece. He has worked in the tech business for many years.
Post reply on HN