"Linux Gazette...making Linux just a little more fun!"


Impressions reading Peter H. Salus `A Quarter Century of UNIX'

By Leif Erlingsson  <leif@lege.com>


I have been involved with Unix and the Internet since '88, and with Linux since '95, but it isn't until reading Peter H. Salus' `A Quarter Century of UNIX' during this summer vacation that I see where Linux fits in into the last 25 years of operating systems development.

Unix came about as a revolt against cumbersome propriety operating systems shipped by the various hardware-vendors. In contrast, Unix was developed by a handful of people. An example of a "huge" software project in the development of Unix is `awk'--developed by three people.

UNICS (original name) was developed at Bell Telephone Laboratories in the Summer 1969 - Fall 1970. Ken Thompson was the initiator and Dennis Ritchie and Rudd Canaday were active contributors.
The intent was to create a pleasant computing environment for themselves. The hope was that others would like it also. The basic notion at the Labs (in Dennis Ritchie's words as quoted from the book),

The Bell Telephone Laboratories staff (BTL) were supposed to discover or invent new things. There was always management encouragement.

It turned out Unix was easy to use and understand when compared to the competition. It was extremely compact. It wasn't until much later that anything and everything the user wanted was supplied (like vi, emacs, X, ksh, csh,... :-)).

The single most important factor behind Unix' popularity was that in the beginning the source code was practically free. Thus it was used in education and as a base for derivate systems. The universities loved it. Later, when AT&T realized that they had in Unix something of great value and tried to capitalize on that, universities were forbidden to use the source code in education. This motivated Andy Tanenbaum to write MINIX, from whence Linus Torvalds got his inspiration to write a kernel for his Intel 386, the kernel that later became Linux.

Bell Telephone Laboratories (50/50 owned by AT&T and Western Electric Company) was, by the so called "consent decree" of Jan 24, 1956 (entered into because of the Sherman Antitrust Act and a complaint filed by the Department of Justice in Jan 14, 1949), required to reveal what patents it held and supply information about them to competitors. Also, the terms of the decree required BTL to license to anyone at nominal fees. So we have this "consent decree" to thank for the phenomenal spread of Unix!

BTL had the following support policy:

This forced the users to band together, which resulted in better and more responsive support than any vendor could have managed. Also, an "us" (users) against "them" (vendors) mentality formed, reinforced by actions taken by AT&T to stifle "the Unix problem".

This is very important: Unix begat Internet!

For a long time no one in business took Unix seriously. For AT&T it was just a legal problem. It was run on VAX'es, but it took the Digital Equipment Corporation about a decade to learn how to support a Unix system as opposed to a Virtual Machine system because of the NIH syndrome. (NIH = Not Invented Here.)

Does it sound like Linux or does it [sound like Linux] ? :-)

On 20 Nov 1974, the U.S. government filed a new antitrust action against AT&T, Western Electric, and Bell Telephone Labs. The settlement reached in 1984 dissolved Western Electric, formed the "Baby Bells" and reorganized AT&T Bell Laboratories into Bell Telephone Labs.

AT&T was now permitted to enter the hardware and software computer business. AT&T sharply raised Unix license fees ...

One reaction was Richard M. Stallman's Free Software Foundation with it's GNU (Gnu is Not Unix) project, that has given the world a wealth of free versions of Unix systems programs. Another is Keith Bostic's CSRG project to create a license free version of Unix. Today, all free Unix clones except Linux use the CSRG code, and all free Unix clones use the GNU code, Linux included.

This is very important: Internet begat GNU and CSRG, and therefore the free Unixes, Linux included. And Unix begat Internet, so therefore, Unix begat Linux. Also, as we all know, Linux is continually developed on the Internet by a looseknit band of programmers from around the world, each doing their little piece -- truly users banded together!

So where do Microsoft and others fit into this picture? DOS/Windows is just one of many systems sprung out of the fountain of Truth -- though there is much debate as to how much truth has rubbed off on them. :-)

There is a huge cultural barrier between the Unix camp and the other guys. It took DEC a decade before the DEC Unix Engineering Group was formed, and when it was, it was located in a separate location from the rest of the company.

Salus tells the story in the book:
there was a lot of animosity towards Unix up and down the company at DEC. Armando Stettner relates how Dave Cutler, one of DEC's engineering elite, at one point got two Unix engineers, Armando Stettner himself and Bill Shannon, to drive down to his office 20 minutes away to help him with, Armando thinks it was, some SRI package on top of VMS. They got there and Cutler was in his office. Armando and Bill sat down at a terminal, and it just didn't do what they expected it to do. Cutler asked them how it was, and Armando replied that it didn't work. To this Cutler said "Well, thank you very much" and they were dismissed. Cutler then called their Senior Group Manager and chewed him out and said Armando and Bill were sorry excuses for engineers and he never wanted to see them in Spitbrook (his office) again. Armando believes that Cutler's disdain has been reflected in his work ever since. Armando says:


To round this off I'd like to itemize a few general factors for the success of Unix:


There must be a fundamental difference of thinking between the free software camp and the other guys:

The first mind-set is to share in order to gain. The other mind-set is hoarding out of fear that something is going to be taken away. Out of the latter mind-set springs the correct business-types managing their various copy-protected products, while from the sharing win-win culture, where each person's efforts becomes a multiplier toward a common goal, springs an open and nonconformistic, somewhat anarchistic type of person. The two often do not like or understand each other.


(This article is copyright Leif Erlingsson. As long as this copyright notice is preserved, and any cuts clearly marked as such, the author hereby gives his consent to any and everybody to use this text.)

(The book `A Quarter Century of UNIX' is Copyright © 1994 by Addison-Wesley Publishing Company, Inc.)


Copyright © 1997, Leif Erlingsson
Published in Issue 21 of the Linux Gazette, September 1997

HOME