Chapter 20. Proliferating Penguins - Part I

From the early 1980s on, the big gripe about Unix was that it had split and resplit, that there were just too many variants. The fact that they had a common base was irrelevant to the critics -- and many (if not most) of those critics were selling VMS or MVS or DOS or...

Following Linus' postings of 1991, there soon were what we have come to call "distributions." And, rather than utilizing ftp, they came on CD-ROM.

The first of these was Adam Richter's Yggdrasil (in the Old Norse Edda, Yggdrasil is the "world ash," from a branch of which Odin/Wotan made his spear). Yggdrasil alpha was released on 8 December 1992. It was called LGX: Linux/GNU/X -- the three components of the system. Recall that Gilmore, Tiemann and Henkel-Wallace formed Cygnus in 1989. Richter spoke to Michael Tiemann about setting up a business, but was "definitely uninterested in joining forces with Cygnus."

Yggdrasil beta was released the next year. Richter's press release read:

The Yggdrasil beta release is the first UNIX(R) clone to include multimedia facilities as part of its base configuration. The beta release also includes X-windows, networking ... an easy installation mechanism, and the ability to run directly from the CD-ROM.

The beta was priced at $50; the production release was $99.

SuSE was formed in 1992 also, as a consulting group (SuSE was originally S.u.S.E., which stood for "Software-und-System-Entwicklung," Software and System Development), but did not release a Linux distribution for several years. The next distribution -- and the oldest still in existence -- was Patrick Volkerding's Slackware, released 16 July 1993, soon after he graduated from Minnesota State University Moorhead. It, in turn, was the basis for SuSE's release "Linux 1.0" of SLS/Slackware in 1994. (SLS was "Softlanding Linux System," Peter McDonald's 1992 distribution, on which parts of Slackware were based.) SuSE later integrated Florian La Roche's Jurix distribution, resulting in a unique distribution: SuSE 4.2 (1996).

The next year, Mark Bolzern was trying to sell a Unix database from Multisoft, a German company. He encountered difficulties because it was relatively expensive to set up the Unix system. Then he came across Gnu/Linux and realized that he now had a real solution. He convinced Multisoft to port Flagship (the db) to Linux and "that was the first commercial product released on Linux," Bolzern said.

"People were always having trouble installing Linux," he continued, "and then Flagship wouldn't run right because something had changed." Bolzern decided that what was needed was a release that wouldn't change for a year, so he "picked a specific distribution of Slackware" and "the name Linux Pro." Soon he was selling more Linux than Flagship: "we're talking hundreds per month."

And when Red Hat came out, Bolzern picked that up.

Mark Ewing had set up Red Hat in 1993. Mark Ewing has said: "I started Red Hat to produce a development tool I thought the world needed. Linux was just becoming available and I used [it] as my development platform. However, I soon found that I was spending more time managing my Linux box than I was developing my software, and I concluded that what the world really needed was a good Linux distribution..."1

In 1993, Bob Young was working for Vernon Computer Rentals. He told me: "I knew the writing was on the wall for my future with that company." He continued:

Red Hat the company was legally incorporated in March of 1993 in Connecticut under the name: ACC Corp. Inc. It changed its name to Red Hat Software, Inc. in early 1996, and changed its name a last time to simply Red Hat, Inc. just before going public in June of 1999.
ACC Corp. Inc. bought the assets, including all copyrights and trademarks (none were registered at the time) relating to Marc Ewing's sole proprietarship business venture in January 1993. Marc's Red Hat project was not incorporated but was run out of Marc's personal checking account. Marc received shares in ACC Corp, Inc. in return for the Red Hat name and assets.

In 1995 Red Hat packaged Linux, some utilities and initial support for $50. Also in 1995, Bryan Sparks (with funding from Ray Noorda, former CEO of Novell) founded Caldera and The Apache Foundation released Apache, which would become the most widespread Web server. But Red Hat soon became the most popular Linux release. This was unexpected: Linus had said that he expected Caldera to be the top supplier, because it was "kind of a step beyond," in that it was targeting the office market. "I think what's interesting about Caldera is they based their stuff on Red Hat and then they added a commercial kind of approach."

When Red Hat became a "success," Bob Young and family moved from Connecticut to North Carolina (Ewing lived in Durham).

It was the end of July 1996. Just in time for Hurricane Fran, the first hurricane to visit Raleigh since hurricane Hazel in 1954. Yes, "the" hurricane Hazel that is the only hurricane to make it to southern Ontario still categorized as a hurricane that I know of.

(Before abandoning this, I should point out that Young is from Hamilton, ON, and attended the University of Toronto. During the night of October 18, 1954, "Hurricane Hazel pelted Toronto with rain and killed 81 people. On one street alone, Raymore Drive, 35 neighbors were drowned." Environment Canada, Canadian Hurricane Centre, Storms of 1954.)

ACC, Young's company, sold Linux/Unix software and books. Young had been introduced to the Unix world in 1988, when he was with Vernon Leasing and Rentals, and began publishing New York UNIX as well as catalog sales. This led to his being the founding editor of Linux Journal, a post he held for two issues in 1994, before "splitting the losses" with Phil Hughes, who is still the publisher of LJ.

In the summer of 1995, I was approached by Lisa Bloch, then the Executive Director of the FSF, as to the feasibility of a conference on "Freely Redistributable Software." I was enthusiastic, but had my qualms about profitability. Richard, at our meeting, was quite understanding: FSF would bankroll the affair, but he hoped we could turn a small profit.

Lisa and I put together a committee (Bob Chassell, Chris Demetriou, John Gilmore, Kirk McKusick, Rich Morin, Eric Raymond, and Vernor Vinge) and we posted a Call for Papers on several newsgroups.

Thanks to "maddog" (Jon Hall), Linus agreed to be a keynote speaker, Stallman was the other. We had a day of tutorials and two days of papers. February 3-5, 1996 at the Cambridge Center Marriott. Everything ran smoothly. By the end, I was a nervous wreck. And the FSF ended up making a tiny profit.


1 See Glyn Moody, Rebel Code (Perseus Publishing, 2001), p. 97.