20 Years Ago in USENIX

USENIX started out as the UNIX Users' Group in May 1974, with two dozen folks in a conference room at Columbia University. As this was meeting 0, USENIX 2K will be the twenty-fifth. Or, at least, that's how I rationalize it. I tried to point this out where the 1985 (Portland, OR) "Tenth Anniversary" was concerned. Clearly we're doomed to be off by one, just like the folks who think the "millennium" begins in January 2000.

Denis the Short (= Dionysius Exiguus) fixed the calendar so that the year after 1 B.C. was 1 A.D. ­ no 0. Of course, the Hindu mathematicians hadn't invented 0 yet, either.

But numbering has been and remains a great problem.

This occurred to me again last Septem-ber. I was giving a talk on UNIX history and was asked a series of numbering questions: What was System IV? Why did Berkeley go BSD, 2BSD, 3BSD, 4BSD, 4.1BSD, 4.2BSD, 4.3BSD, 4.4BSD? (I'm ignoring 1a, 1b, 1c, Tahoe, etc.)

I've never known what System IV was. Perhaps AT&T decided that as they had System III and there was a 4BSD, it would be too confusing to have a IV. If anyone knows the reality, please let me know.

The Berkeley "scheme" was a bit more straightforward: the licensing agreement negotiated with AT&T permitted upgrading the software and supplying bug fixes, but barred a major release without renegotiation. So, Bill Joy and his fellows at the CSRG decided that by indicating 4.0, 4.1, 4.1c, . . ., this would show that they were "updating," not producing a "major revision." Hmmm.

And that's the way the rhinoceros got his skin.