Though his company publishes some of the best Linux books out there, Tim O'Reilly says Apache, not Linux, is the battle ground for the open source revolution.
O’Reilly & Associates’ familiar zoo of helpful books have been our best friends during many a long hard night spent coding. But there is more to Tim O’Reilly than books alone – he has been a driving force behind the commercialization of free software.
|all photos © gary wagner |
It’s actually a pretty simple thing to get to O’Reilly & Associates headquarters from San Francisco. All you have to do is turn your car away from Silicon Valley and drive. It’s appropriate that a man who has made a name for himself identifying (and profiting from) trends just outside of the mainstream of the computer industry would choose to locate his company in the more out-of-the-way locale of Sebastopol, CA rather than in the heart of the Valley. Since he started producing Unix manuals in 1983, Tim O’Reilly has been on the cutting edge of the Internet and open source revolutions. In fact, his company started the Internet’s first Web portal (Global News Navigator), sold the first ad on the Web, produced the first commercial Web server, and helped organize the branding campaign for Open Source software. And O’Reilly publishes the Perl book with the camel on it. He recently invited Linux Magazine Publisher Adam Goodman and Executive Editor Robert McMillan for a tour of his Sebastopol offices and a chat about the evolving open source industry.
Linux Magazine: How did you first get into Unix?
Tim O’Reilly: At the very beginning I was a contract tech writer. The first book that I developed on my own was actually an early version of Unix in a Nutshell. I was working at a company called MassCom that’s now long gone.
LM: Weren’t you mowing lawns at some point?
TOR: I was a cemetery caretaker at Arlington National Cemetery. That was when I was in high school.
At MassCom, I had been hired to write these manuals on their proprietary graphics stuff. I finished the manuals and I started learning Unix. And I began to ask people there, “How are your customers going to figure out all this Unix stuff?” The people at MassCom said, “You’re right. We need a manual.” So I wrote what was probably the first Unix system administration manual. I didn’t know squat about Unix. I just went around, interviewed their Unix guys, and figured it out.
LM: What year was that?
TOR: That was probably 1983. Unix was written by a group of people who weren’t thinking of it as a commercial system. At that time everybody began talking about the fact that Unix was this industry-standard operating system that all these companies were beginning to use. But what’s really ironic is that they were all writing their own documentation from scratch because there was no good documentation available. I thought, “If the software is going to become standard, we’re also going to have to have a standard for the documentation, so everyone doesn’t have to do their own.” That was where we fit in. Basically we were a consulting company that wrote whatever people wanted us to write.
LM: How do you stay on top of emerging technologies? Do you do this consciously or is this just an extension of your position in the community?
TOR: It’s a mix. We certainly have individual editors who track a specific technology area, but a lot of it is over the transom. How did we get into Perl? It was totally random. One of our system administrators was into Perl, and she got some e-mail from Randal Schwartz saying, “Hey do you think O’Reilly would be interested in a book on Perl?”
LM: Was that Perl book your all-time bestseller?
TOR: No. Our all-time bestseller was probably the The Whole Internet User’s Guide and Catalog. In terms of impact, too, that was probably our most important book ever. It really had a major social impact because it was the first introduction to the Internet for people outside the technical community; many people first learned about the Internet from that book.
GNN [Global Network Navigator] was a spin-out of The Whole Internet User’s Guide. We tried to put the catalog part online, and it became the Web’s first important site and the first Web site to have advertising. Now this book has been out-sold by the true mass-market publishers. More people have read Internet for Dummies or DOS for Dummies than any of our best-selling books, but I don’t think these books have had the same impact.
LM: What happened with GNN? Do you think it was a mistake to sell it to America Online? They never really did anything with it, but GNN was basically Yahoo!, a few years before Yahoo! even existed.
TOR: I have to say, I’ve wrestled with that a lot over the years. Obviously I made a mistake in some sense, but there’s no guarantee that GNN would have become Yahoo!. Yahoo! might well have beaten our socks off anyway. But we were always really outside the money system. Yahoo! hooked up with the venture capitalists and they became part of this wonderful system that spawns great companies. We didn’t do that; we’re kind of a homegrown outfit. I didn’t want to play that game, because to me the stakes were pretty high. You know, for every Yahoo! that succeeds, a lot of companies have failed. Maybe they have even enriched their owners. But I was trying to do something different. My goal was to build a company that would still be here a long time from now. And since I already had a successful business, I felt that I didn’t need the money from outside. I did not maximize the opportunity in that sense.
The only way I’ve had regrets is that I felt like, gosh, I could have made a lot of people who worked with me rich, assuming we’d succeeded.
LM: It’s really interesting to hear you say that, given what’s happening with Linux these days. As the financial community has become interested in open source, many Linux companies are facing these same kinds of questions. What did you feel you’d have to be giving up if you played the VC game? Was it that you felt investors would have too much control over the company?
TOR: It was fear. It was not knowing what would go on, and also not realizing the stakes. I mean obviously none of us knew how big it would get. If it had been a choice between taking the risk of creating a multibillion-dollar company or losing it, I might have made the choice. But at the time, it looked like I’d risk giving it up for something that might not be more substantial than my existing business.
LM: What prompted you to get involved with Linux?
TOR: It was through our German company and Sebastian Hetze, who wrote what was probably the first Linux book in Germany. Sebastian had self-published his book, and we tried to persuade him to publish it with us. That got us interested.
There’s actually something curious about Linux. We did some early Linux books and they never did all that well. And we’ve never quite figured this out, even to this day. Except for this recent real upsurge, Linux is not that important a book market, relative to its supposed importance. It just isn’t. It’s really odd. I mean, with a book like Running Linux or Linux in a Nutshell, we’ll probably sell 50 or 60 thousand copies a year. This is substantial, but it’s not on the order of magnitude you would expect from a community with 15-20 million users.
Some of this might be a fashion thing in technology. With certain technologies, people really need information. With others, they don’t. I would imagine some of this is because a lot of Linux people were Unix people and they’re simply bringing their skills over. They don’t really need it. And there was a lot of good online documentation already available.
Actually, I like to remind people that I still don’t think that Linux is the most important open source technology out there.
LM: What is?
TOR: The Internet. People don’t give it credit, but the Internet is a product of the open source community in more ways than you can count. It’s funny because Rick Adams, who founded UUNet, will say, “I don’t think open source had any impact.” Rick Adams wrote “B” news. It was the most widely used Usenet news software. He wrote the most widely used implementation of SLIP [Serial Line Internet Protocol], which brought the Internet the last mile. And he started the first ISP. So here’s this whole multibillion-dollar ISP industry, which was started by somebody from the free software community, building a business to support the culture that had grown up out of this network community.
Now look at the whole way the Internet standards evolve. It’s very much a collaborative community activity, much like Linux. People have written this before, but if you look at the IETF [Internet Engineering Task Force -- the volunteer body that defines Internet standards], it’s just like Linux. Anybody can show up, anybody can participate in it. You make technical contributions. No, it doesn’t have the Free Software rhetoric around it, but as part of a cultural phenomenon, it’s really the same thing.
LM: What about the argument that the Internet is a creation of the platforms upon which it rests, and to the extent that Microsoft can control the platform via Windows, Linux is a more important technology because it maintains an open foundation for the Internet?
TOR: Well, Linux is politically more important. It’s an important battle front that we have to win. The point is precisely because the open source community wasn’t self-conscious at the time that we lost control of the Internet. I joined the board of the Internet Society (which is sort of the parent, legal umbrella over the IETF) because I saw this happening. Back in the mid-1990s I saw that we had to get across the idea the Internet belongs to its users, or else it was going to end up belonging to companies. In fact, that’s pretty much what’s happened. So all my open source activism is actually my second wave of community-oriented activism because I do not want that to happen again.
I have to say, I don’t necessarily just consider Microsoft the bad guy. Being realistic you know, Netscape started the battle of proprietizing the Internet. They really had an explicit strategy of trying to seize control of Internet protocols, and of course Microsoft jumped on that bandwagon and beat them at it.
I think there was a time when we should have drawn the line a little differently. Maybe we couldn’t have. Maybe Microsoft would have done it anyway, but I certainly think, even now, there’s a lot more opportunity to have a voice that says the users matter, and I think that’s part of what I see as the significance of both the Internet and of open source. It’s a shift of power and it’s a new medium that shifts power in a whole lot of ways that we don’t understand yet.
In the long term, I feel the focus of the Linux community on trying to beat Microsoft on that platform is irrelevant. Look at where the interesting applications are being developed. The browser was the last really significant desktop application. All the new applications are now being delivered in the browser. At the end of the day, what’s the new application? Amazon, E*TRADE, eBay, MapQuest. That’s where the new functionality is happening. It’s happening on the Internet and they are basically on this new paradigm.
LM:But isn’t it true that if you control the browser, the interface to the Internet, you can then dictate the standards?
TOR: Yes and no. You could take that kind of argument down and say: well, Linux isn’t really free because you’ve got to write to this proprietary hardware architecture — the Intel processor. That’s bogus. If at the end of the day Microsoft controls both ends of the connection, that is a problem. Because then they really can control what you can and can’t do.
I think Apache plays an enormously important role here. Because it has dominant market share, it keeps the Internet open. I think it’s more important for Apache to have dominant market share than for Linux. If Linux is dominant too, that’s better, but I’d hate to see us lose Apache. That’s a really important battleground.
LM: So what will be Linux’s legacy?
TOR: It’s really hard to know what Linux’s legacy will be because it’s too early to tell. It will probably keep Microsoft on its toes. They’re going to have to deliver some new value, and so, at the very least, the challenge will probably give Microsoft a little more life in their dead platform.
At best, we’ll be a credible challenge and we’ll basically break open some new standards. I hate to say this because the Linux community doesn’t like to hear me say it, but I just don’t think that operating systems matter as much as people think they do.
LM: Recently you applauded Sun’s Community License efforts as a new approach to open source, but some people see Sun’s licensing of Jini under its Community Source License as a cynical attempt to control this emerging Internet platform. How do you reconcile this?
TOR: I’m not saying I like everything about the Sun license. In fact, I’d prefer an Open Source license. Sun is one of these companies that is saying: “If we can get some of these crucial choke points, we can be in the driver’s seat.” Absolutely. But the open source story is complex.
I had a meeting recently with some people from Sun about their license. I said: “What’re you trying to do? Are you trying to energize a community of developers? Are you trying to energize the Linux community to work on your stuff? If you’re just trying to energize your own developers, does it work or does it not?”
In the ideal world, maybe they would be more open, but I don’t think we have any right in the open source community to try to tell people: “Thou shalt do X,Y, or Z.” It’s either better for them or it’s not. Now, if it’s better for them, what are the axes on which you can say it’s better?
Some people — Richard Stallman for example — have a philosophical motivation. Sun has an economic motivation. They’re a public company; they’re trying to maximize shareholder value and so they are sitting there saying: “What’s the right balance between doing the right thing and making money?”
What we want to do is figure out a way to shift the balance in favor of more open rather than less. We know that Sun is not going to ever go all the way over. The question is how far can you get them to go?
I say let them experiment, and one of two things will happen. Maybe they will find out that they don’t get a lot of community participation this way. Then we’ll get to say to them: “Look, you guys, you missed something. The Sun Community Source License did not work.”
On the other hand, if it does work and it does enable some participation, what we’ll have done is we’ll have gotten the whole computer industry to move a little bit more toward a developer-centered world in which they recognize the power of individual developers.
Then we can move on to the next island. I think a half-open development model is a heck of a lot better than a closed one. I’m not willing to say: “Either go all the way or we hate you.” It just seems silly to me.
LM: What do you think of Red Hat’s business model?
TOR: To me, the market reaction to Red Hat is like the market reaction to a lot of Internet companies. There are a lot of people who don’t really understand what’s going on but think that this could be the next big thing. What is it they say about investment? Stocks are bought in hope, held in greed, and sold in fear.
So the market reaction to Red Hat and the business of Red Hat probably don’t have a lot of connections. Like they don’t for a lot of Internet companies, because it’s too early to tell what the real impact will be. I feel that if Red Hat succeeds, it will probably be in some unexpected way. At the end of the day, this is what VCs understand.
I see a lot of parallels with the PC industry. When it started out, people said, “We can’t take this too seriously because there’s not that much demand for computers. Why would you need one?” Some people had the vision to answer this question. Actually, one of those visionary people was Bill Gates. He believed that there would be a computer in every home, and he drove that vision.
LM: Do you believe there is a market for competing versions of Linux?
TOR: Absolutely. There are definitely different systems. I mean, I think there’ll be a shakeout. There’s no guarantee that any of the current Linux vendors will be the winner. I think it’s pretty clear that Linux and other open source platforms are going to matter in a tremendous way.
Look at the early history of the PC. Who were the major companies? IBM and Compaq are the only ones to survive. When you first bought commodity PCs, there were a dozen companies that are all out of business now. Then a second generation came along — some of them were big companies getting into this new market. I would be very surprised, for example, if we don’t see a Microsoft/Linux distribution on the shelves one day.
TOR: Yes. It makes a lot of sense to me. Microsoft may say, “Never,” but there are a lot of companies that had based their business on their proprietary computer architectures who ended up deciding to ship PCs. At some point, there’ll be an economic imperative that says: “There’s a slice of the market that’s looking for this particular kind of functionality.” Just as IBM sells mainframes, mid-range computers, and PCs, Microsoft will build some kind of Linux open source product line.
I’ve told Microsoft directly: “Open source is good for you.” Even with their business model they’ve made a lot of money because of these independent hackers coming up with innovations that they can use.
LM: Right now a lot of people from the Microsoft world are looking at Linux. But the traditional Linux world and the Microsoft world seem very polarized.
Do you think that they will meet at some point? Do you think that there are lessons that the Linux community can learn from Microsoft? Do you think the Linux community needs to, for example, acknowledge that there will be users that will not really care about technology?
TOR: The vast majority of users simply do not care. As long as Linux depends on users caring, it will not penetrate that vast hinterland. This is just like the Web. I think Linux will win, not because it re-creates the Microsoft world on the Linux platform, but because it builds a new world.
The Web is even easier to use than all those point-and-click interfaces and the fact is, apart from a few fairly simple operations, the kinds of things that were delivered as part of traditional point-and-click operating systems, don’t matter that much anymore. Literally.
I think it’s extremely unlikely that Linux will displace Windows directly by a frontal assault on the desktop or desktop applications. I think it’s going to come around the back way by supporting new applications — by being that Intel inside the next generation of computer applications.
Robert McMillan is executive editor for Linux Magazine. He can be reached at firstname.lastname@example.org.