Open Source's First Six Months

Eric S. Raymond

Issue #55, November 1998

If anybody had suggested to me then that the paper was going to motivate something like the Netscape source release, I would have wondered what they'd been smoking.

Back in March, Netscape announced their intention to release the source code of Navigator. Since that time, we've seen once again that very few things are as powerful as an idea whose time has come.

I'm reminded of this every time I surf the Web. The Open Source meme is everywhere. It seems you can't open a technical or business magazine these days without tripping over an admiring article about Linux—or an interview with Linus Torvalds—or an interview with...er...me.

Half by accident, I've ended up near the center of all the crazy and wonderful things now happening. When I composed “The Cathedral and the Bazaar” a little over a year ago, I was aiming to explain the Linux culture to itself and to explore some interesting and somewhat heterodox ideas about software development. If anybody had suggested to me then that the paper was going to motivate something like the Netscape source release, I would have wondered what they'd been smoking.

But that's exactly what happened, and I soon found myself in the role of leading advocate and semi-official speaker-to-journalists for a hacker community suddenly feeling its oats. I decided to take that job seriously, because somebody needed to do it and I knew how and nobody else was really trying very hard. I had the advantage of experience; I'd been in this role before, for lesser stakes, after the New Hacker's Dictionary came out in 1991.

The point of all this personal stuff is that I've had an almost uniquely privileged view of the early days of the open-source revolution—as an observer, a theorist, a communicator and an active player in helping shape some of the major events.

We've come a long way, baby...

In this essay, I intend to do three things. One, celebrate the incredible victories of the last few months. Two, share my thinking about the battles being fought right now. And three, consider where we need to go in the future and what we need to do, to ensure that open source is not a mere fad but a genuine transformative revolution that will change the rules of the software industry forever.

When you're living on Internet time, it can be hard to remember last week, let alone last year. Take a moment and think back to New Year's Day, 1998. Before the Netscape announcement. Before Corel. Before IBM got behind Apache. Before Oracle and Informix and Interbase announced they'd be porting their flagship database projects to Linux. We've come a long way, baby!

In fact, we've come an astonishingly long way in a short time. Six months ago, “free software” was barely a blip on the radar screens of the computer trade press and the corporate world—and what they thought they knew, they didn't like. Today, “Open Source” is a hot topic not just in the trade press but in the most influential of the business newsmagazines that help shape corporate thinking. The article in The Economist in July was a milestone; another was the August issue of Forbes with an explanation of the concept as their cover story and a picture of Linus on the cover.

The campaign also went after corporate endorsement of open-source software. We've got it, in spades. IBM—IBM!—is in our corner now. The symbolism and the substance of that fact alone is astounding. [Apache is the web server shipped with their Web Sphere product.]

We haven't shot ourselves in the foot...

The last six months are also notable for some things I had feared would happen, but did not. Despite initially sharp debate and continuing objections in some quarters, the hacker community did not get bogged down in a loud and divisive factional fight over the new tactics and terminology. Bruce Perens and I and the other front-line participants in the Open Source campaign did not get publically savaged for trying to gently lead the community in a new direction. No one burnt us in effigy for actually succeeding.

The maturity and pragmatism with which the community backed our play made a critical difference. It has meant that the story stayed positive. We have been able to present open source as the product of a coherent and effective engineering tradition, one able to sustain the momentum and meet the challenge of what the corporate world considers “real support”. It has denied the would-be bashers and Gates worshipers among the press the easy option to dismiss us all as a bunch of fractious flakes.

We've all done well. We've gotten our message out and we've kept our own house in order—and all this while continuing to crank out key advances that undermine the case for closed software and increase our leverage, such as Kaffe 1.0. What comes next?

Toward World Domination

I see several challenges before us.

First: The press campaign isn't over by any means. When I first conceived it back in February, I already knew where I wanted to see positive stories about open source: the Wall Street Journal, The Economist, Forbes, Barron and the New York Times.

Why those? Because if we truly desire world domination, we must alter the consciousness of the corporate elite. That means we need to co-opt the media that shape decision-making at the highest corporate levels of Fortune 500 companies. Personally, all the press interviews and stuff I've done have been aimed toward the one goal of becoming visible enough to those guys that they would come to us wanting to know the Open Source community's story.

This has begun to happen (besides the Forbes interview, I was a background source for The Economist coverage)--but it's nowhere near finished. It won't be finished until they have all gotten and spread the message, and the superior reliability/quality/cost advantages of open source have become common knowledge among the CEOs, CTOs and CIOs who read them.

Second: When I first wrote my analysis of business models, one of my conclusions was that we'd have our best short-term chances of converting established “name” vendors by pushing the clear advantages of widget frosting. Therefore, my master plan included concerted attempts to persuade hardware makers to open up their software.

Although my personal approaches to a couple of vendors were unsuccessful, Mr. Eid Eid's (then president of Corel Computer) speech at UniForum made it clear that CatB and the Netscape example had tipped them over the edge. Subsequently, Leonard Zubkoff scored big working from the inside with Adaptec. So, we know this path can be fruitful.

A lot more evangelizing remains to be done. Any of you who work with vendors of network cards, graphics cards, disk controllers and other peripherals should be helping us push from the inside. Write Bruce Perens (bruce@pixar.com) or me (esr@thyrsus.com) if you think you might be positioned to help; combination Mister-Inside/Mister-Outside approaches are known to work well.

Third: The Interbase/Informix/Oracle announcements and SGI's official backing for Samba open up another front. Actually, we're ahead of my projections here; I wasn't expecting the big database vendors to roll over for another three months or so. That third front is the ability to get open-source software into large corporate networks and data centers and in roles outside of its traditional territory in Internet services and development.

One of the biggest roadblocks in our way came from people who said “okay, so maybe Linux is technically better, but we can't get real enterprise applications for it.” Well, somehow I don't think we'll be hearing that song anymore. The big database announcements should put the “no real applications” shibboleth permanently to rest.

Our next challenge is to actually get some Fortune 500 companies to switch over from NT to Linux or *BSD-based enterprise servers for their critical corporate databases and to go public about doing it.

Getting them to switch shouldn't be very hard, given the reliability level of NT. Waving a copy of John Kirch's white paper (http://www.kirch.net/unix-nt.html) at a techie might be sufficient. In fact, I expect this will begin to happen swiftly even without any nudging from us.

However, that is only half the battle. Because the ugly political reality is this: the techies with day-to-day operational responsibility who are doing the actual switching are quite likely to feel pressure to hide the switch from their NT-leaning bosses. Samba is a huge win for these beleaguered techies; it enables open-source fans to stealth their Linux boxes so they look like Microsoft servers that somehow miraculously work well.

There's a problem with this, however, that's almost serious enough to make me wish Samba didn't exist. While stealthing open-source boxes will solve a lot of individual problems, it won't give us what we need to counteract the attack marketing and FUD-mongering (fear, uncertainty and doubt) that we'll start seeing big-time (count on it) as soon as Microsoft wakes up to the magnitude of the threat we actually pose. It is not enough to have a presence; we need a visible presence—visibly succeeding.

I have a challenge for anyone reading this with a job in a Fortune 500 data center: start laying the groundwork now. Pass the Kirch paper around to your colleagues and bosses. Start whatever process you need to get an Oracle- or Informix- or Interbase-over-Linux pilot approved—or get prepared to just go ahead and do it on the “forgiveness is easier than permission” principle. Some of these vendors say they're planning to offer cheap evaluation copies; grab them and go!

I and other front-line participants in the Open Source campaign will be doing our best to smooth your path, while working the media to help convince your boss that everyone's doing it, and that it is a safe, soft option that will look good on their performance report. This, of course, will be a self-fulfilling prophecy.

Fourth: Finally, of course, there's the battle for the desktop—Linus' original focus in the master plan for world domination.

Yes, we still need to take the desktop. The most fundamental thing we need for that is a zero-administration desktop environment. Either GNOME or KDE will give us most of that; the other must-have, for the typical non-techie user, is an absolutely painless setup of Ethernet, SLIP and PPP connections.

Beyond that, we need a rock-solid office suite, integrated with the winning environment, including the “Big Three” applications—spreadsheet, light-duty database and a word processor. I guess Applixware and StarOffice come close, but neither are GNOME- or KDE-aware yet. Corel's port of WordPerfect will certainly help.

Beyond repeating these obvious things, there's not much else I'll say about this, because there's little the Open Source campaign can do to remedy the problem directly. Everyone knows that native office applications, well-documented and usable by non-techies, are among the few things we're still missing. Looking around Sunsite, I'd say there might be a couple of promising candidates out there, like Maxwell, a WYSIWYG word processor, and Xxl, a powerful spreadsheet. What they mainly need, I'd guess, is documentation and testing. Would somebody with technical writing experience please volunteer?

We're winning!

Yes, we're winning. We're on a roll. The Linux user base is doubling every year. The big software vendors are being forced to take notice by their customers. Datapro even says Linux gets the best overall satisfaction ratings from managers and directors of information systems in large organizations. I guess that means not all of them are pointy-haired bosses.

The explosive growth of the Internet and the staggering complexity of modern software development have clearly revealed the fatal weaknesses of the closed source model. The people who get paid big bucks to worry about these things for Fortune 500 companies have understood for awhile that something is deeply wrong with the conventional development process. They've seen the problem become acute as the complexity of software requirements has escalated, but they've been unable to imagine any alternative.

We are offering an alternative. I believe this is why the Open Source campaign has been able to make such remarkable progress in changing the terms of debate over the last few months. It's because we're moving into a conceptual vacuum with a simple but powerful demonstration—that hierarchy, closure and secrecy are weak, losing strategies in a complex and rapidly changing environment. The rising complexity of software requirements has reached a level such that only open source and peer review have any hope of being effective tactics in the future.

The article in The Economist was titled “Revenge of the Hackers”, and that's appropriate—because we are now remaking the software industry in the image of the hacker culture. We are proving every day that we are the people with the drive and vision to lead the software industry into the next century.

Eric S. Raymond can be reached at esr@thyrsus.com.