|
20 Years Ago in USENIX
by Peter H. Salus
June 1979. Toronto, Ontario. It's really tough trying to write this up. For example, Brian Kernighan wrote me: My memory is quite dim. According to my calendar, I gave a talk at the software tools group in Toronto, and it says that I gave a "v7 talk" at "UNIX user's group," but I have no memory of the latter and hardly any of the former. David Tilbrook wrote: Toronto firsts: Rob Pike and David Tilbrook give maiden USENIX speeches together on QED. And Rob wrote: That was a very long time ago. I don't remember anything much, although it might come back to me with prodding. I do recall the discussion afterwards included comments from the audience about ex/vi, and there was an exchange about the high cost of screen editors, which interrupted the operating system for every keystroke. Thanks to Geoff Collyer (who must throw out even less than I do), I have the program(me)s for STUG and for USENIX, as well as his notes. STUG, the Software Tools Users Group, met on Tuesday, June 19. Kernighan made some introductory remarks (including the news that UNIX/RT "is not likely to be released for five years"). He was followed by Dave Hanson (then of the University of Arizona) on portable file and I/O systems and Doug Comer (then as now at Purdue) on his "Mouse4," a rewritten preprocessor that used a hash table. After a break, several of the Georgia Tech group gave a paper on their tools running on a PRIME. (My view of their work is in A Quarter Century of UNIX, chapter 12.) Dennis Hall then delivered a status report on LBL's Virtual Shell. He said: My strongest memory is that we weren't really legitimate. I felt like a pretender. Actually, I never quite got over that feeling. The true blue UNIX people thought we were wasting our time on second best. The true blue VMS (or any Brand X) users thought we were debasing their systems. Remember the person who thought it would be a good idea to build a TSO shell interface for UNIX using the tools? After lunch, Robert Munn (University of Maryland) gave a status report on his ratfor preprocessor, which was being used to distribute a crystallography course to 250 students. The afternoon concluded with a "Haves and Wants" session and a discussion of distribution and standards led by Kernighan. The USENIX conference proper began on Wednesday with a session on languages (Pascal, Euclid, and YASL). After lunch, there was a session on UNIX, including papers by Brian Kernighan (V7) and Tom London (V32). Thursday morning began with a talk by Al Arms of Western Electric, who announced the horrendous new fee structure: $20K for V6, $30K for PWB, $28K for V7, and $40K for V32—per CPU! He also said that it was forbidden to teach the internals of V7. Not a good showing from the corporate set. There was a discussion of the user group and of the various compilers available from Whitesmiths'. After lunch it was text processing and graphics (Dennis Murmaugh, Martin Tuori, Tom Duff, and Bill Reeves). Tom showed a videotape of computer-generated images (512 x 512 x n; where n=8 or n=24). Geoff Collyer recalls that "it was better resolution than TV." The afternoon session concluded (as Tilbrook recalled): "Bill Buxton gives concert . . . probably first musical presentations at a USENIX or any other computer conference." Collyer wrote that it was "very boring." In the evening there was a reversi (=Othello) tourney. Friday began with a session on mini UNIX, MRS database, realtime UNIX (with papers by Neil Groundwater and by Eric Ostrum), networking, and several applications: an RT-11 emulator, an accounting system, STUG tools, and Idris. The afternoon session was on system improvements: multiple-address space, supporting 64 terminals and 1,000 users on an 11/70 (Ian Johnstone), large UNIX systems (George Goble), high-performance UNIX (Mike Muuss), and porting UNIX to a Univac (Hale Pierson). Saturday morning was sparsely attended. (Most folks had been drinking from the wine-and-cheese party early on till the wee hours.) Whoever won got the reversi award. Tilbrook and Pike gave the first of many QED papers (this one called "the ultimate version"). Bill Joy then spoke about what was going on at Berkeley, and Wozniak about what was going on at BBN. Most of the attendees knew what was transpiring at Berkeley, for Bill had distributed a four-page document, "Second Distribution of Berkeley Software for UNIX." You could get a 2,400-foot 800BPI tape for $60. The tape contained the Pascal system, "a new version of the 'ex' editor," "a new shell 'csh,'" a new troff macro package (-me), and a new mail program. I'll let Dave Yost have the last word: I witnessed BSD history being made quite casually at the Toronto USENIX that year. Bill Joy was standing in the lobby at the rear of the auditorium when Jim Kulp from IIASA (in Laxenburg, Austria) approached him to pitch "Job Control." He had it all working and wanted to offer it to Berkeley for inclusion in the distribution. I was excited; it sounded great. Bill, whose job it was to receive such submissions and then do a lot of the work integrating them, was noncommittal but said he'd look at it. The rest is history.
Postscript
Hi, I'm one of Brian [Hudson]'s "15-year-olds" ("20 Years Ago in U[SE]NIX," ;login:, April 1999), though now of course I'm 35. I was quite pleased to see Brian and those of us lucky enough to be his students at the time mentioned in your column. Brian taught us enough about C and about UNIX internals and system administration so that we students could run and enhance the system on our own. This system (an 11/70) supported activities such as class scheduling and student records, including our grades, and our school newspaper was laid out using our Diablo daisy-wheel printer. This in addition to our favorite activities: extending the system, writing and enjoying games, etc. We had an awful lot of fun back then. We learned a lot of good and useful skills. We had the pleasure of seeing the results of that work, and the trust and feeling of community that developed.
Farrell Woods
|
 Need help? Use our Contacts page.
Last changed: 7 Oct. 1999 mc |
|
Issue index