@sicp @ThatWouldBeTelling

> I figure makes the most sense if you consider writing stuff line-by-line on a teletype with ed; it keeps the lines short.

This is a myth, it's a facile explanation at least as old as the UHH. "The only reason to name it `cp` instead of `copy-file-from-one-location-to-the-other` is because of the balky teletype!" Even if it weren't a myth in the case of Unix, it fails to explain APL, J, K, Forth. Even Paul Graham noted that terseness is useful. "The C Programming Language" by Kernighan and Ritchie, the book that the "K&R" style is named after, was a book, it was intended to be read, not executed. I/O redirection on Multics, at the time Bell Labs was still involved, looked like this:

iocall attach user_output file filename
list
iocall attach user_output syn user_i/o

...Where Unix just uses `ls>filename`. Which one's easier to read? Once you know what ">" means, obviously "iocall attach user_output file" is *harder* to read. You catch the difference between "file" and "syn"? Those lines are identical for the first 26 characters: sifting out signal from noise has a nonzero cost.

TUHS has the v6 source, the v7 source, the v10 even. Glass tty by the time v7 rolled around. The Plan 9 source is all over the place: bitmap displays with mice and everything. A lot of the code reads like Dick Gabriel (WIB author) described: "I'm always delighted by the light touch and stillness of early programming languages. Not much text; a lot gets done. Old programs read like quiet conversations between a well-spoken research worker and a well-studied mechanical colleague, not as a debate with a compiler. Who'd have guessed sophistication bought such noise?"

Here's something: the length of a string of digits that a person can keep in their short-term memory is proportional to the length of time it takes them to pronounce those digits. People whose native languages use less time to pronounce the numbers can remember longer strings of numbers; genetics not involved, just the language you grew up speaking.

Code is language; it's an artificial language, but it's a language. Think about mathematics: word problems in school textbooks, Roman numerals, prose descriptions of equations versus equations written in the standard-ish algebraic notation we use nowadays. The conventional wisdom, that "more characters" means "easier to read" is completely fucked. Pike's example in "Notes on Programming in C" ( http://doc.cat-v.org/bell_labs/pikestyle ) was "MAX" versus "MAXIMUM". Using more words than you need to doesn't make your writing any clearer. Prose and math and phone numbers and people think that it's somehow the *opposite* for code.

And so the shell, or any of the code, you think more than you type. You describe these programs to people: you type them or you say them. You think in your head, "I had a clever solution to this, what did I type yesterday?"