Linus Torvalds - Just for Fun

Now there would be two big-name open source projects -- Netscape and Linux -- and the reasoning was that if Netscape, the better known of the two, were to fail, the reputation would impact Linux, too.

And to a large degree, Netscape did fail. The company had trouble getting open source developers interested in the project for the longest time. It was just a huge body of code and the only people who could get into that code were Netscape people.

The project was somewhat doomed not only because of its size but also, because Netscape wasn't able to make everything available as open source -- only the development version, which was fairly broken at the time the company released it. The company couldn't GPL the browser because not all the code was theirs -- the Java portion was licensed from Sun, for example. Not everyone on the newsgroups agreed with Netscape's license. On the whole, the license was fairly mellow, but if you're someone like Richard Stallman you don't like mellow.

I thought it was wonderful that Netscape took this step, but I didn't view it as a personal achievement. I remember that Eric Raymond took it really personally. He was extremely happy about it. His paper, "The Cathedral and the Bazaar," which did an excellent job of explaining the open source philosophy and history, had been released the year before and was cited as one of the reasons behind the Netscape decision. He was actively pushing open source. He had been at Netscape on a number of occasions, trying to convince them to open up their browser. I was there only once. In fact, Eric had visited a number of companies bearing the open source message. I was interested in the technology, not the evangelization.

Within 24 hours of Mozilla being released, an Australian team that called itself the Mozilla Crypto Group created the cryptography module. Back then, non-U.S. citizens were prohibited from using encryption generated on U.S. soil. Suddenly, somebody from Australia had done the work, so non-U.S. citizens were in a position to use it. But there was a catch. Given the export restrictions of the time, the Mozilla project couldn't take the Australian code. If it made its way to the United States, it couldn't be reexported. This meant that one of the first successes of the great Netscape experiment couldn't become part of Mozilla.

We were all worried because Netscape had received a lot of news coverage. And for that first year, people walked on eggshells. Nobody wanted to say anything negative about Netscape for fear that it would result in bad press for open source and scare other companies away.

But two months after Nerscape's move, Sun Microsystems joined the game by declaring that it would become the first major hardware vendor to join Linux International. It would support Linux on its servers. The company with the unimpressive licensing scheme for its Jini project had decided that Linux was worth taking seriously. The newsgroups overflowed with self-congratulations. With Sun on board, Linux developments made their way from Internet discussions to the trade press. Outsiders were suddenly interested, but mostly technical outsiders.

Then came IBM.

IBM has been known for being fairly stodgy, so everybody was taken by surprise when the company announced in June that it would sell and support Apache, the most popular commercial Linux version used for Web servers.You could run Apache on AIX, IBM's UNIX, and that's probably what a lot of people who bought IBM did. That's how Apache got to IBM's attention. Somebody must have noticed that most of those server machines ended up using Apache, so they calculated that they would sell more servers if they had the in-house expertise to support such customers. Or maybe they were acting on feedback from customers who said they would buy IBM machines but would run Apache.

It's relatively easy to install Linux on a computer. But for most companies, one of the big issues, historically, has been: Who do we blame when something goes wrong? Obviously, there are the Linux companies like Red Hat that provide the support, but it was a psychological advantage for customers to know that IBM would be there for them. When IBM started getting into open source, a lot of people suspected it was just lip service. But that turned out not to be the case. IBM dipped its toes in the water by running and supporting Linux on its server boxes and then sort of waded all the way in. Next came the small PC servers. Then, the regular PCs. Then, the laptops. The company has announced it will spend $1 billion on Linux this year.

IBM did a lot of its Linux work on its own. I think one of the reasons they liked Linux was because they could just do what they wanted to do without having to deal with licensing issues. It's a company that has had its share of hassles. IBM was screwed over by Microsoft after the two companies jointly developed the OS/2 operating system, which turned out to be just Windows on steroids. Microsoft failed to support OS/2 because it wasn't interested in sharing the market. Windows NT is what came out of it from the Microsoft side. But OS/2 never paid back to IBM the billions of dollars poured into it. And IBM was plagued with the licensing issues over Java. I think they were just happy not to have all that aggravation with Linux.

There's no doubt that IBM was Linux's biggest coup. And it generated only excitement on the newsgroups -- not the sort of paranoia provoked by the Netscape announcement, or any of the seething anticommercialism that has periodically (okay: frequently) divided Linux enthusiasts.

By July, Informix announced that it would port its databases to Linux, meaning that even if you used Linux to operate your computer, you could run an Informix database. It wasn't such a big deal at the time. The company had been having financial trouble, but it was still one of the top three database vendors. Linux people were mildly happy about the development, and were writing self-congratulatory essays in Linux advocacy groups.

Within weeks, from out of nowhere, Oracle followed suit. Oracle dominated databases. Long before the announcement there had been rumors (on the newsgroups) about the company having some internal ports to Linux. And, since Oracle is synonymous with Unix servers, it wasn't such a major leap to Linux. But if you followed the newsgroups, we had definitely entered the big time. The Oracle announcement had a huge psychological impact, even if its technical impact was zero.

Like the IBM announcement before it, Oracle's big move was felt not only by the Linux community but by the folks who are commonly referred to as management decision-makers, although some people prefer the term "suits." No longer would they be able to say that they couldn't use Linux because their business depends on databases.

While the news was gratifying, it didn't change my life. Tove and I were juggling two adorable kids. Most of my nonfamily hours were spent on Linux maintenance, both at home and in the office. To keep from favoring anyone version of Linux, I used Red Hat at work and SuSE,a European version, at home. At one point I felt I wasn't getting enough exercise, so I decided to ride my bicycle the six miles between our apartment and Transmeta's headquarters.It was on a Monday. There were no hills to climb, but a strong wind blew in the wrong direction, making it more challenging than I wanted. By the time I left work ten hours later, the wind had shifted so that it was still in the wrong direction. I phoned Tove and she picked me up. Needless to say, biking-to-work didn't happen again.

I add this innocuous detail only to illustrate that the Linux developments weren't affecting my daily life. Most of the activity was taking place at corporations. Technical people, who had long known about Linux were being approached by their companies' leaders who had been seeing articles about Linux in the trade press, or hearing about it. They would ask their technical folks what the fuss was all about. Then, once they learned the benefits, they would make the decision to have their servers run Linux.