20 Years Ago in USENIX

USENIX Historian

The USENIX Association held a conference in Boulder, CO, from January 29 through February 1, 1980. It was preceded by a meeting of STUG, the Software Tools User Group. It was also preceded by the appearance (after 18 months!) of ;login:, vol. 5, #1.

STUG was largely concerned with its forthcoming distribution tape and its contents. But two other things proved really important.

First, recall that Kernighan and Plauger published Software Tools in 1976. Andy Tanenbaum introduced the LBL crowd (Hall, Scherrer, Sventek) to it, and they were enamored and set to writing tools an a virtual operating system. Also enamored were the students at Georgia Tech. They also worked on tools, using their PDP-11, but they were aiming at PrimeOS.

Second, at the end of the STUG session, Debbie Scherrer announced that the LBL group was looking for someone "to run our UNIX systems." In the audience was a young man attending his first USENIX meeting. He says he literally ran to the front of the room, "throwing chairs out of my way." The group took him out for Chinese lunch—interview, and so Mike O'Dell left the University of Oklahoma for Berkeley, where he became their UNIX guru and the ARPANET liaison. (The LBL 11/70 was an early ARPANET host.)

I've my copy of "GIT-ICS-79/07: Georgia Tech Software Tools Subsystem User's Guide September 1979" sitting here as I write. It stands as a monument to Perry Flinn, Allan Akin, and Dan Forsythe, who wrote the contents:

   Subsystem tutorial
   PRIMOS File System Overview
   Software tools text editor
   User's Guide for the Command Interpreter
   User's Guide to the Ratfor preprocessor

and

   Software Tools Text Formatter User's Guide

Half a dozen years later, Dan Forsythe was one of the organizers of the Atlanta USENIX (June 1986); Debbie Scherrer served on the USENIX Board for many years, including terms as president and vice president; O'Dell also served on the board and was the founding editor-in-chief of Computing Systems.

Stay with me a while, you'll see what I'm getting at.

Al Arms was at Boulder, too, on behalf of Western Electric. He informed the 450 attendees that the Justice Department had said that the UNIX licensing agreements are "compatible" with the consent decree. He also announced a new small-systems license at $700/user to $9,400 for an (unspecified) larger number of users.

Bill Joy spoke about his work on implementing VAX/UNIX paging.

Lou Katz, president of USENIX, announced that he expected the distribution tapes to begin going out around April 1.

Tapes were featured in ;login:, too:

   Fourth Software Distribution
   Submissions for the Fourth Software Distribution may be brought to the Boulder
   meeting or mailed to arrive in New York before February 15, 1980. On that date we will
   start packaging the distribution with a target date for first mailings of April Fool's Day . . .

The issue of ;login: also contained copies of the Articles of Association and the bylaws of the Association.

The entire issue of ;login: (as well as summary notes on the Boulder meeting by Ian Jackson [U. of Sydney]) appeared in the Dec.—Jan. issue of the AUUG Newsletter.

In 1955, during the transition from the IBM701 to the 704, a number of "operators" in California got together to share software and hardware fixes.

With IBM's encouragement, this grew into SHARE.

Till the late 1970s, source code wasn't a question: code came with your machine. Those brown 8-inch floppies from DEC!

With the advent of USENIX began the wholesale exchange of hardware and software bug fixes and — in 1976 — the swapping of tapes. (The First Distribution was May—June 1976, the Second in November.)

Just to give you a "taste," the Second Distribution contained contributions from the RAND Corporation, the Naval Postgraduate School, UCSD, Yale, and the University of Illinois. The Third followed in May 1977 (when USENIX held its meeting in Urbana, IL, with 150 attendees).

It's hard to express just how important these tapes were — they contained software from all over that became indispensable to the users. USENIX continued distributing tapes into the late '80s, when distribution by ftp made them unnecessary.

However, here's the nub: open source is far older than Linux or GNU. Accessible source and shared code have been with us as long as we have had real computers. In fact, the IBM 701 and 704 were large machines with thousands of diodes and triodes. And, at a point where the ARPANET had under 100 hosts, those USENIX distribution tapes were the way to get the stuff around.

The way you get versatile, robust code is by letting everybody poke at it. Open code is better code.