author, The Cathedral and the Bazaar
O'Reilly & Associates
Are big corporations learning the lessons of the Open Source movement? We recently talked with Eric Raymond, author of The Cathedral and the Bazaar, and Tim O'Reilly about corporate America's deepening infatuation with open source technologies and methodologies.
Listen to this interview:
Eric Raymond: I think it's a harbinger of several important trends. One is that IBM, like many other software producers, is recognizing the increased reliability and performance and other desirable features that you get from decentralized peer review. Economically I think it means IBM has also figured out that operating systems is a bad business to be in these days.
Tim O'Reilly: Yeah, I always loved Bob Young's remark that he was really trying to shrink the size of the operating system market. And I think, in fact, that that's just one of the key concepts that I think Open Source has really brought into the frame: Software is becoming a smaller part of the overall picture, and you've got to somehow have a strategy that goes beyond just selling your bits.
Raymond: That's right, and there are two ways to look at that, one of which looked kind of negative at first blush, and the other one of which looks kind of positive. The negative way to look at it from a traditional business person's and investor's point of view is that we're squeezing the leverage out of the software market. There's less and less room for investors to play by pumping in capital if the development networks are all decentralized and open source. On the other hand, this also implies that the investors have really been getting their leverage from inefficiencies in the conventional mode of production, and when those inefficiencies are squeezed out, everybody will benefit.
O'Reilly: With Collab.net, which is the company we started with Brian Behlendorf, which is aimed really at that marketplace of trying to help corporations embrace the methodology and use it, we're finding really good uptake on that. People are understanding that this is what they wanted, they just haven't known how to do it. I guess I see a lot of growth coming from companies that -- and this goes back to Eric's thing about inefficiencies in the market -- there's so many cases where the customers of these guys are reinventing the wheel, and the open source methodology allows the customers to sort of get on the same plane and start sharing data.
|"Secrecy, and the kinds of closed intellectual properties that are bound up with secrecy, are not efficient." -- Eric Raymond|
Raymond: That's right, and very directly what we're discovering is that secrecy, and the kinds of closed intellectual properties that are bound up with secrecy, are not efficient. They're not an effective way to generate value in software.
O'Reilly: You know the thing that I guess I see, a lot of people ... What metaphor do you use? Different metaphors in open source are better for explaining different parts of the problem. And so, for example, I see the first wave of the open source enthusiasm was a little bit driven by one aspect of the Linux story, which was, "Oh, look, they give it away for free and everybody uses it and isn't that a wonderful network effect?" So all these people came along and said, "Oh, we're going to give away our software for free, we'll make it open source and bingo everybody will use it."
Raymond: That's right, but the meaning of the long-term future of this revolution is going to be driven not by price but by value.
|"The Apache lesson is ... the users basically formed a collaborative network to keep the power on their side of the fence." -- Tim O'Reilly|
O'Reilly: Absolutely. But to go further along, what I'm seeing in terms of businesses, you know beyond that first wave of people who thought, "Oh we can just jump easily on the bandwagon," we're starting to see people for example who are trying to learn the Apache lesson instead. And I think that's a really fascinating lesson for business.
The Apache lesson was, here's a bunch of end-users, ISPs who basically were being disempowered by their vendors. You know, they were using a free solution, all of a sudden Netscape and Microsoft started saying, "No, no, the main path going forward is our proprietary solutions," and the users basically formed a collaborative network to keep the power on their side of the fence. And I guess I see a lot of opportunities where businesses are -- vertical markets are really cooperating to build some technologies they use to run their business. And that's what Eric talked about in this whole idea of "use value of software" rather than "sale value of software."
Raymond: When companies invest in a shared infrastructure, they can reap cost-efficiencies because they're not paying for the whole infrastructure if there's a whole consortium like the Apache group that's being funded by multiple corporations. Every time somebody joins that gang, the costs for everybody go down and benefits go up.
Sims: It reminds me actually, one of the earliest interesting things about the Internet itself was its ability to route around damage, and I made a big deal about that in the early days of the Internet, that it would route around damage. That's sort of a metaphor for what you were talking about. The damage was they couldn't get what they wanted out of the vendors; they routed around it. It's almost a sort of monopoly fail-safe in a way.
Raymond: Yeah. I like Gilmore's quote about the Internet interpreting censorship as damage and routing around it. These days I like to generalize it to: The Internet interprets attempts at proprietary control as damage and routes around it.
O'Reilly: We all know that hardware margins have come way down, but software margins have stayed where they are because Microsoft had this proprietary lock-in. And now everybody's saying, you know, "We want you guys to enjoy the same kinds of margins the rest of us have to live with."
Raymond: Yeah, actually I think that's going to turn around and bite Microsoft very badly in the near future. I'm giving a section of my talk nowadays called "The Seven Bullets Microsoft Has to Dodge to Survive the Next 18 Months," and I think of those the most difficult for them to dodge is the basic contradiction in their business model between the fact that their revenues have to rise every quarter to sustain their option program, and the fact that hardware prices are plummeting. That puts an intolerable margin squeeze on their hardware OEM standard. I think we're mirroring the point where those OEMs are going to revolt.
Sims: If I can change tracks just a moment here, we've been talking about the motivation of open source, and how the methodology works, and part of the methodology of that -- even Microsoft identified this a couple years back in its "Halloween Documents" -- was the fact that they weren't just, coders weren't just in it for the profit. They were in it for a couple things. Some of them wanted to make the software better. Some of them wanted the recognition, and what that actually might lead to down the road. But I wonder, especially in light of the aftermath of V.A. Linux's IPO, I wonder how wealthy coders, wealthy open source coders, can change that equation. I mean does that change the equation in anybody's view of how open source methodology works or does it have no impact? Does it actually free people up to explore new technical challenges?
|"I don't think all this money coming in is going to make a major difference to the development model or the dynamics of the community." -- Eric Raymond|
Raymond: Well, your listeners deserve to know that I'm on the board of directors of V.A. Linux, and in theory I made a lot of money from that IPO, so they should listen to what I say knowing that and taking it into account however they like. My personal view is that I don't think all this money coming in is going to make a major difference to the development model or the dynamics of the community. And the reason I'm confident about that is because demand for programmers has been intense for a very long time, for over a decade. I think that all the people who could be seriously distracted by money are already gone.
O'Reilly: I have to disagree a little bit, Eric. One of the things that I certainly see is that it's certainly true on the individual programmer level, but if you look at projects as competing for mind share, there's definitely an issue whereas these big companies need to maintain their market cap and maintain their momentum, I think there's going to be a little bit more dog-eat-dog in the open source corporate space than we've seen. And that's going to lead to some bad blood.
Raymond: Well, you might be right but I'm not sure that that's relevant to the question he asked which is "Do we expect this money to change the behavior of developers?"
O'Reilly: Well I think it does if, for example, you're working for a large open source company and that company says, "We're going to do this project because we're basically trying to out-do the other large open source company or out-do this startup that might be threatening our future revenue stream, so we're going to basically rip them off." You know, you've got companies that are effectively redirecting the efforts of those people who still think of themselves as open source programmers but are really now actually working on strategic corporate directives which may not be motivated by the same kinds of open source behavior.
Raymond: Well, you may be right. On the other hand, when I look at open source projects that have been sponsored by these newly successfully large companies, it looks to me like the developers are more wedded to their projects than they are to their employers. There are a couple of cases in which, when key developers have moved to another company, they've actually taken their projects with them.
O'Reilly: Yeah, I agree in the sort of part where the projects themselves are open source, but when you look at how various Linux companies are hoping to make their revenue, they are definitely looking to steal the wind out of the sails of other open source companies, and there's definitely some behavior that I'm starting to see that I think is going to give the open source community a bad name. I'm not going to mention any names, but ...
Sims: Companies aggressively seeking open source development, people seeking open source technologies ...
O'Reilly: No, just where they're basically, they see somebody developing something interesting and they say, "Oh, we need to do that because if that's going to be a good revenue stream, we better have that before the other guy does," and you know it's becoming more competitive and more dog-eat-dog underneath the rhetoric of "We're all idealistic and sharing."
Raymond: Well, Tim, when I see evidence of that happening, I'll probably issue some kind of broadside about it, but I haven't seen it yet.
Sims: Speaking of people who are secure financially and turn instead to a technical challenge, Bill Gates's decision to move to chief software architect from the CEO position. I'm just wondering if this, at the highest level, reflects any sort of industry-wide move -- recognition that the technical challenges are at least what drive developers at heart.
O'Reilly: I don't know, I have to say it's certainly true to say that Bill Gates has moved beyond money. I'm not sure what's motivating him at this point, but I don't think that his move has much of anything to do with the stated story of "I'm going to do interesting technical challenges." It's a little bit of a shell game, you know, in the Justice Department face-off where he's basically trying to make himself a little bit less of a target.
Raymond: I kind of wonder, actually, if he isn't getting out of the line before the stock price crashes.
O'Reilly: Oh, well that's an interesting thought. So in other words, if there's a dump, he can say "Not on my watch" and preserve his reputation for posterity.
Sims: When we were talking about corporations embracing open source, Sun's decision to open Solaris 8, I wonder what either of you thinks of that.
Raymond: I think it's a shuck. What they're doing is not really opening the source. It's creating a sort of pseudo-openness where they talk the right game but keep proprietary control of the results.
O'Reilly: Let me make a slightly othogonal point to that, Eric. I've actually encouraged, not directly with Sun, I'd rather see them go a full open source route. But at least in theory, I think it's great that they're doing something like this because we will learn so much from it. There are so many different aspects of the open source story, and part of it has to do with, well, how much of it is simply just having access to the source code so that you can fix your own bug. You know, we hear that as one of the benefits. Well, under the Solaris model you'll be able to do that. So then the next question is, "Well, okay, no, no there's this whole part about redistribution, so that you can ... " Well, that part they're not doing. So what we'll find out is which of those things is more important to the community. We'll have a lot more data about what really matters in open source.
Raymond: We will find that out. I have confident expectations of what the answer will be.
O'Reilly: Yeah, but I'd rather find it out empirically, so I'm glad that Sun is running the experiment. Although I have to say, I mean, it's not a pure experiment in the sense that, you know, Linux has so much momentum and at bottom, this just looks like kind of a halfway me-too gesture which I think, at bottom, will just reinforce the Linux momentum.
Raymond: That's my read on it too.
O'Reilly: There's probably a time when Sun could have done it and reaped more benefits.
Raymond: If they had done this two years ago, it would have been big news. Now it's not.
O'Reilly: The other thing when I talked with Sun, I also pointed out you know it's kind of like break-up value in companies. You know, they asked me this three or four months ago, "Do you think we should open source Solaris?" and I said, "Well, even if you went full open source with it, I think you'd get more mileage out of open-sourcing pieces of it than you would out of open sourcing the whole thing." You know, the story's over. People don't want another open source operating system. You know, they want to improve the ones they've already got. So, for example, open source NFS will figure out what parts of their code base are better and put some of those things out, they would have got a lot more leverage than trying to unload the whole thing as a competitor to Linux.
Sims: That will allow people to redistribute that and make it ...
O'Reilly: Sure. Scott's always had this idea of, you know, "over my dead body will do X," and so they always come up with ...
|"There's been a pattern of corporate announcements of 'No way will we have anything to do with open source.'" -- Eric Raymond|
Raymond: Yeah, there's been a pattern of corporate announcements of "No way will we have anything to do with open source," and then two weeks later, pretty reliably, you see some kind of open source announcement. I interpret that those statements are a sign that something internal may be about to crack.
O'Reilly: Yeah, I think so too. I'd be surprised if we don't see some more, I won't say legitimate, but more real open source announcements from Sun, and at some point, they're basically looking at all this. They're basically market testing, if you like, different approaches to see how people react, and I think the momentum is certainly there for open source and for people to say, "Okay, you know, you either go all the way or don't pretend."
Sims: Well, given what you just said, Eric, is it unrealistic to expect an MS-Linux?
|"I am fairly sure that there is already, however, a Linux-portable Office. I have some intelligence from inside Microsoft that strongly suggests that." -- Eric Raymond|
Raymond: That is not something I feel like I have a good answer to. I am fairly sure that there is already, however, a Linux-portable Office. I have some intelligence from inside Microsoft that strongly suggests that, and it also makes sense for that to exist already if the people at Microsoft are smart enough to see that there's a wreck coming in their operating systems business -- and I think they are that smart.
O'Reilly: Yeah, I mean, you don't want to lose both markets.
Raymond: My projection of the immediate short-term future at this point is I think the OS monopoly is going to break decisively some time in the first six months of 2001, with the defection of one of the major gray-box OEMs, somebody on the level of a Dell or a Gateway. And within a week after that happens, we're going to see an announcement of Office for Linux.
Copyright © 2009 O'Reilly Media, Inc.