November 8, 2005

Interview: Bob Young after Red Hat

Author: Joe 'Zonker' Brockmeier

Bob Young is, arguably, one of the most influential figures in the development of Linux and open source. By co-founding Red Hat with Mark Ewing in 1993, Young helped turn Linux into a household name. After being involved with Red Hat for more than 12 years, Young recently stepped down from Red Hat's board of directors. We caught up with him to see what his plans are, and what his thoughts are on Red Hat and the future of open source.

Young is now spending his time focusing on his most recent business venture, Lulu.com, a print on demand service where content creators can sell their books, comics, movies, or any other content that can be digitized and sold over the Web. The company handles sales and distribution of the work, and takes a cut from the sale price as well as a charge for the media itself. Users can also purchase copies of their work to sell directly.

How did Young make the leap from Linux to self-publishing? Lulu.com actually has its roots in the short-lived Center for the Public Domain (CPD), a non-profit Young founded with Mark Ewing in 1999. The CPD's mission was to help combat the expansion of intellectual property laws that were, as Young put it, "the biggest single threat to the open source movement."

The CPD shut its doors in 2002, and Young expressed dissatisfaction with working in the not-for-profit space. According to Young, what led to Lulu.com was the idea that the expansion of intellectual property laws was tied to the consolidation in the publishing industry, and wanting to solve the problem from a business point of view.

"You know, there really is a problem associated with the expansion of intellectual property, and if in a free market democracy, the citizen and the consumer are the same person, there's got to be a way of solving the problem in the marketplace."

Still, one might wonder why Young chose now to exit Red Hat. Young said that the work he was doing on Red Hat's board, while important, wasn't his cup of tea because the "bulk of your time is spent on the minutiae of dealing with SEC regulations and you're not really spending that much time helping the executive team fashion strategy; you're spending most of your time doing defensive things for the company, making sure that the company stays on side of all these arcane rules." At the same time, Young said that Lulu.com is "booming" and he wanted to devote more time to "helping Lulu get to the next stage."

Lulu isn't doing quite as well as Red Hat just yet, but the company is growing at a respectable pace. Young said that the company's sales are increasing at a rate of more than 10% a month. "We're selling in excess of 35,000 books a month. I believe we sold 40,000 in September. We have well over 30,000 books in our library, so to speak, of books. Gives you an idea that if you grow those numbers 10% a month, it doesn't take too many years before this is a very big business."

Open source makes it possible

Just because Young is no longer with Red Hat, it doesn't mean that he's lost his passion for open source. Young said that Lulu wouldn't be possible without open source software. "It is safe to say it would have cost well in excess of twice the investment, and it's cost several million dollars' worth of investment to get to Lulu where it is today. And it would have cost twice as much as that if we'd had to use proprietary software to do all of that."

In addition to cost savings, Young said that having access to source code has allowed the company to move more quickly. "Because when you're a startup, innovation is everything. Whatever you decide to do on day one is not going to work -- I mean, that's just a given. So therefore you have to be able to adapt your offering very quickly, going forward in response to your customers. And you can't do that if you're using Oracle databases and proprietary Web servers and proprietary programming tools.

Bob Young of Lulu.com (Photo by M.J. Sharp)

"It improves your ability to build better tools, because you're not guessing at what's going on inside the database or inside the Web server. You're actually looking at the code, and when you build your application on top of it, you're not guessing at whether you're doing the right thing. You can actually watch the code cycle through and you can go, 'Jeez, now I see why it hiccupped on line 56.' You're not guessing at it anymore and that's simply a dramatically better way of building tools, especially building Internet-based tools, than trying to do it on proprietary software where you actually don't know what's going on."

Lulu.com contributes to some of the open source projects that it uses in the course of its business. The company has contributed to the development of the content management framework bitweaver, as well as LJBook, a program that converts blogs in a LiveJournal format into a PDF document that's ready to print. (Lulu.com uses PDF for sending books to the printer.)

Looking back at Red Hat

Now that Young has closed the door on his involvement with Red Hat, we asked if there was anything he would have done differently, given the benefit of hindsight. Young's response is that "you don't look gift horses in the mouth -- the point being that as an entrepreneur, to start a business in your wife's sewing closet and to have it be valued by the international financial market at over $3 billion 10 years later, you kind of go, 'No.' That's one of those size of success, quantity of success that you simply can't second guess."

Young had mentioned that he originally encouraged his engineers to use Fedora Core rather than Red Hat Enterprise Linux (RHEL), but that the company was now standardized on RHEL 4. Now that Young is on the customer's side, we asked what he thought of Red Hat's pricing model, which has drawn more than a little criticism as the company has grown. Young said that he thought Red Hat was approaching it the right way.

"Red Hat's offerings have to get better. They are not nearly good enough yet. And the people who are demanding these better offerings are major Fortune 1000 corporations around the world. So for the startups, or the Lulus in 2002, our Red Hat made Fedora available so all of the code Red Hat distributes, they distribute under an open source license, and as a consumer of open source and Linux technology, that's all I would ask them to do. How they then come up with a business model that enables them to pay for the engineering that major Fortune 1000 corporation need is a separate issue, and one that's between Red Hat and its customers."

We also asked Young about the rumors that Red Hat CEO Matthew Szulik had met with Microsoft's Steve Ballmer. If Young were CEO, would he be willing to sit down with Microsoft? Young said he couldn't comment on whether Szulik had sat down with Microsoft, but that he would be willing to talk to Microsoft. "You talk to everyone. As my father points out, talk is cheap. But the key way to think of it is unless you sit down with everyone, you will not be properly educated."

On the GPL and the future

Earlier this year, Eric Raymond ruffled a few feathers by saying, "We don't need the GPL anymore." We asked Young if he too thought that the GPL was no longer necessary. Young said that he "couldn't disagree more," though he was quick to point out that he is simply "disagreeing with an interpretation of a friend of mine ... Eric and I are approaching this from different perspectives."

Because everyone is familiar with the GPL, Young said, there's no need for developers to read the license or worry about someone sneaking something in the back door. "So I would strongly argue just the reverse: that we will never outgrow the GPL. And that the GPL will continue to be the dominant open source licensing model, not because the GPL is the best license, but because it is the standard license."

What about Linux on the desktop? If Young's habits are any indication, it will be a while. Young's desktop of choice is Apple's Mac OS X. "I could not bring myself to go to Windows, of course. And one of the little-known facts -- 'cause, you know, I'm a skinny guy who wears glasses, everyone assumes that I'm an engineer by training -- I'm an old typewriter salesman by training. So maintaining my own Linux desktop was just something that I didn't have the training to do." He did note, however, that Lulu.com's engineers are all using Linux on their desktops.

Looking forward, we asked Young what other areas in the tech industry he found exciting. Young said that he still found the Internet exciting, though many people think that "there's no real opportunity for the little guy anymore.

"My take on it is just the reverse. It's that the big guys on the Internet today are gonna look small compared to the big guys on the Internet 10 years from now. And that there will be household names on the Internet 10 years from now that you and I have yet to hear of today."

Read the entire transcript of the interview ...

NewsForge.com: Let's start off, just talk a little bit about Lulu.com and what kind of made you decide to go in that direction?

Bob Young: Well, it's interesting enough. There is actually a connection between Lulu and my open source Red Hat days and it's simply this: Lulu was really started to -- well let me give you the full background. When I left Red Hat or was sort of easing my way out, Mark Ewing and I started a thing called the Center for the Public Domain. Because what we're doing is in '99 when, you know, Red Hat had that big public offering, we had a debate of what can we do -- Mark and I -- for the open source community that had done so much for us. And we looked at, you know, all the software projects we might have supported and we realized then every open source software project you could name was getting support at that time.

So we took a look at it from the opposite point of view of "If the open source movement failed, why would it fail?" And our conclusion there was that our legislators, literally around the world, in Brussels and Washington and Tokyo, had all come to the conclusion that, you know, obviously technology is an important industry and they come to a big conclusion that intellectual property, more intellectual property, meant a healthier technology industry. And while I'm a big fan of both copyrights and patents, the problem was that our legislators didn't recognize the fundamental rule, which is: too much of a good thing no longer is. And so we're seeing things like the DMCA, like the idea that you could patent ideas, not just inventions, like the idea of taking copyright from 20 years to a hundred years with very little public debate on the topic and you sort of realize that it's a little bit like vitamin D -- you know, too little vitamin D and you get a variety of health problems. Too much vitamin D will actually kill you, I don't know if you know that one. It's actually a poison when taken in large quantities.

And you know, that's what all this intellectual -- this rush to expand intellectual property rules without any serious public debate looked like to us. And it was, without a doubt, the biggest single threat to the open source movement. There were legislations -- legislative initiatives -- that would have materially harmed the open source community's ability to work collaboratively on software.

So that's what the Center for the Public Domain was, and you know, after running a non-profit for two years, I learned everything I didn't want to know about the not-for-profit business, which is that it's very hard to define what success is so when I was looking at this whole issue from a business point of view, you know, an Adam Smith capitalist's view of the idea that enlightened entrepreneurs working on their own self-interests will create as much value to society as the most enlightened monarch. And going, "You know, there's got to be -- there really is a problem associated with the expansion of intellectual property and if in a free market democracy, the citizen and the consumer are the same person, well then there's got to be a way of solving the problem in the marketplace. And that's really what led to Lulu, was our conclusion was the expansion of intellectual property with what -- we thought there was a correlation between the consolidation we're seeing in publishing everywhere with the expansion of these intellectual property rules and so we said, "If you're seeing all these consolidation in publishing, who's being harmed?"

The people being harmed are the authors because with fewer publishers, then there's fewer books being published. And the consumers, with fewer books being published, there's less choice. So at this point, Economics 101 textbooks, there's pretty much exactly six because there are six major textbook publishers. Each of them has an Economics 101 textbook and none of them need a second Economics 101 textbook. So there's six Economics 101 textbooks and there's got to be 5,000 people -- professors of economics in North America alone who could've written a perfectly good Economics 101 textbook. So from a consumer point of view, why are there only six if there's 5,000 people capable of writing one and the answer is, because of the consolidation in the publishing industries and this is happening, of course, across the publishing spectrum.

NF: Now, one of the arguments kind of like when Red Hat and other Linux vendors started getting into the business, Microsoft and other companies objected on, 'Oh, it's open source so it won't be as good a quality.' You hear professional publishers claiming that without that filter, you're going to get a bunch of crap, basically.

BY: Yup. It's sort of the, I don't know, irony or perhaps the -- I guess it's sort of an irony. Here you have business people who don't seem to have any confidence in the power of the marketplace. You know, so the logic that you have to have someone in control of the content and to ensure quality leads you back to a very Marxist dictatorship-type approach. We really should have a government deciding what is good for us because we're not capable of deciding for ourselves. And so the publishing industry, to say that, you know, sort of having -- empowering -- a whole new generation of authors to bring their work directly to the marketplace, is somehow not in the interest of consumers, is very similar to the logic that the old Russian Stalinists used to use on why, you know, the state should run the economy.

NF: How's Lulu doing so far?

BY: It's doing remarkably well, which is actually what led to my stepping away from my final responsibilities at Red Hat because as a director of a public company these days -- with Sarbanes-Oxley and everything else -- the bulk of your time is spent on the minutiae of dealing with SEC regulations and you're not really spending that much time helping the executive team fashion strategy; you're spending most of your time doing defensive things for the company, making sure that the company stays on side of all these arcane rules.

And so I was just looking at it, saying, "That's it; this is really high value work but it's not work that I personally bring a lot of value to." And so Lulu, on the other hand is booming and I needed to find some more hours to devote to helping Lulu get to the next stage. So it's sort of an easy decision of saying, "Look, I'm doing role value work for Red Hat that they can find many more other talented directors to do for them. And I need to apply more hours to Lulu."

So, Lulu at this point, we've been for more than a year, 14, 15 months now, we've been growing at faster than 10% per month, month over month. And we only have to look at that trend line and even while Bill Keiser, a fellow board member at Red Hat and a leading VC, one of the guys who helped Red Hat go public back in '98, '99. When I pointed this to him, he, you know, he chuckled and said, "Bob, I think you're benefiting from the law of small numbers." Meaning, you know, you start with a dollar worth of sales and it isn't actually that difficult to take it to a $1.10 worth of sales.

NF: (Laughs) Right!

BY: And it's not that difficult to take that to, you know, a $1.21 worth of sales. You know, this sort of thing. Having said that, John Maynard Keynes would say the miracle of compound interest implies that, you know, with the volumes that we are now doing and we're still seeing this growth rate carry on. We're selling in excess of 35,000 books a month. I believe we sold 40,000 in September. We have well over 30,000 books in our library, so to speak, of books. Gives you an idea that if you grow those numbers 10% a month, it doesn't take too many years before this is a very big business.

NF: How extensively is Lulu using open source?

BY: We would not -- Lulu would not be possible without open source software. So the two sort of connections back to, you know, my training in the open source world, is one understanding that if you see an industry that is, you know, whose economics are all about advancing the profits for the industry and who are not as focused on the consumer -- and the software industry looked very much like that, the proprietary software industry where all the engineers wanted source code, but none of the suppliers of software would give their customers the source code. And you just go, "Hold on, that's not right." That's not how, you know, a free market's industries are supposed to work. You're supposed to do what their customers want them to do, not they want to do.

So in publishing, the same phenomenon is occurring where the publishers are going up, instead of coming up with new and better books and new and better services. Look at the music publishing space, where the music publishers instead of innovating, they wait for Apple to do the innovations and their contribution is to hire a bunch of lawyers and go around and sue their customers. And any time you see an industry start react to change by suing customers, you go, "Holy cow, there's got to be an opportunity there." So that's kind of where we're going with Lulu, is really to try and address that, so the two connections are: one is we're trying to empower consumers to be able to get their hands on the books they want as opposed to the books that a narrow leash of established suppliers are willing to offer them. But the other one is open source software.

Lulu would not be possible without all the open source tools we used. It is safe to say it would have cost well in excess of twice the investment and it's cost several million dollars' worth of investment to get to Lulu where it is today. And it would have cost twice as much as that if we'd had to use proprietary software to do all of that. Because when you're a startup, innovation is everything. Whatever you decide to do on day one is not going to work -- I mean, that's just a given. So therefore you have to be able to adapt your offering very quickly, going forward in response to your customers. And you can't do that if you're using Oracle databases and proprietary Web servers and proprietary programming tools.

Because at the end of the day, your vendor has control over what you can and cannot offer your customers. Because you don't have source code, you can't add functionality to those pieces of software. So it has done two things for us. One, it has dramatically, and as I say that a minimum of 50% savings on the cost of the software in terms of investment cost; but secondly, it has enabled us to innovate, to evolve our application forward dramatically faster than we could have done had we not access to the source code and had we not had a license that allowed us to modify the tools that we are using. Lulu's primarily based, it's a PostgreSQL database-based tool. We use the Apache Web servers. We started with Fedora but we are now on RHEL 4. We became a customer of Red Hat not because I told anyone to, in fact I was encouraging them to stay on Fedora. We became a customer for all the reasons that Red Hat insists that you should become a customer, which is that it's reduced our sys admin cost dramatically by using the Red Hat tools.

NF: Looking back on the time that you spent with Red Hat, with the benefit of 20-20 hindsight, is there anything you would have done differently in running the company or decisions you would have made differently?

BY: No. (Laughs) It stems from the very big picture issue of, you know, you don't look gift horses in the mouth, you know. It's sort of when you have a success of that scale, to go around second guessing the success and say, "Wow, instead of creating a billion dollar company, we really should have created a $2 billion company." Actually, I think Red Hat is valued right now in excess of $3 billion.

The point being as an entrepreneur, to start a business in your wife's sewing closet and to have it be valued by the international financial market at over $3 billion ten years later, you kind of go, "No." That's one of those size of success, quantity of success that you simply can't second guess. But even more than that, actually even at the time when we were still working early in this thing and for that matter, still to today, sometimes you just follow the stars and the sky, you know, there's this path that you just keep doing what seems obvious. And it turns out that the obvious thing is exactly the right thing to do.

So one of the bizarre things of the whole project was just how -- but of course this is true for, you know, I'm sure the guys at eBay would, you know, tell their story the same way as the guys at Google; but that once you understand whether you're lucky and you stumble into it or whether you're really smart and you engineered the opportunity, once you find an opportunity that other people don't see, you really drive towards that opportunity.

More on page 3...

NF: Are there any, since Red Hat obviously is, you know, kind of a dominant Linux player right now, are there any things, what do you think Red Hat did differently that allowed it to become the primary market leader, as opposed to, say, Caldera before they became evil, or Slackware or something like that?

BY: Two key things I guess they did is: One, Red Hat never -- actually, three, when I think about it. I have to keep track of these things. One is that we focused on the big picture: opportunity. We were never competing with the other Linux distributions.

Two, would be engineering capability that Red Hat built; that's largely Marc Ewing and Eric Troan's contribution. And then the third one was our commitment to giving our customers control over the software we're asking them to use. So coming back to these, the first big picture was, we'd still to this day don't think of ourselves as very successful because our competition was never with Slackware. We always liked Pat Volkerding and what he was doing with Slackware; you know, he was an ally of ours. As are the guys at Debian, as are the guys at SUSE.

Our competition, for better or worse -- and it used to be highly intimidating, you know. I'd be working away in my wife's sewing closet, Marc was down in his wife's and his spare bedroom in their little apartment in Durham, and yet, you know -- when I would talk to my wife and she'd say, well, you know, "Bob, this seems like a high risk venture. You know, if something goes wrong, what'll happen?" And I'll go, "Well, $44 billion a year Microsoft will roll over in their sleep and squash us like a bug." (Laughs) She says, you know, when you look at it from that point of view, that when we started Red Hat all of our competition, every single person making an operating system was billing in the billions of dollars, you know.

See, you had Microsoft, you had IBM, you had Apple -- I guess the smallest guys were SCO at the time and they were doing a $100 million a year worth of business. And Marc and I were, you know, trying to do this out of our spare bedrooms, you know, who were we kidding? Except that we were able to do the one thing, this is the third of my list. We're willing and able to do one thing for our customers that our much, much bigger competitors simply weren't prepared to do; which is that we're willing to give our customers control over the software that we're asking them to use. And no one else was willing to do it and when you would talk to, you know, the engineers who were building the Web sites, you know, supporting the Internet, all of them had to have source code if they were going to enable their servers to talk to everyone else's servers as reliably as they needed to be able to work in order for the Internet to work properly.

So, you know, we are all very focused on things like Linux and Apache as successful open source projects, or maybe MySQL or Postgres. The reality is probably the very biggest open source project in terms of its impact on the world is the domain name serving server tools. Because had those not been open source, whoever owned those would have owned the Internet. But also had they not been open source, the reliability of exchanging messages around the Internet would have never gotten to the level that it is today.

NF: What do you think about the moves by Massachusetts to basically require open document formats and other governments that are legislating open source at various levels?

BY: Finally, some of these guys get it. And you knew this was going to happen because even our legislators, you know, we keep turning over legislators and if you're the average age of your, whatever, member of your state, house, or even member of congress, is -- pick an age -- fifty, it means 10 years ago, they were trained in computers 20 or 30 years before that, so in the 1970s. But we turn over our legislators so today, they're still 50 years old but they were actually trained only 20 years ago when you know, the first ideas of the Internet, and of open source, and the PC revolution came along. And another 10 years from now, our legislators will have learned computing, when Linux and open source and Apache were beginning to take over the world.

So what you're seeing in Massachusetts is very practical stuff. You know, this guys are looking at their budgets. They're trying to get the maximum value for the tax dollar, cause their constituents sure don't want to pay more taxes, and meanwhile these guys are looking at their budgets and they're saying, "We're writing these huge checks to these huge proprietary software companies in California. Is there a better way of buying technology?" And you know, when they do that analysis, they very quickly recognize that if they can control over the technology they are using, they can -- suddenly they can go rout for support, they can bid their support workout instead of being forced to go back to manufacturer of the software or of the file format. And the moment you can do open bidding, it reduces your cost dramatically.

NF: Microsoft has been making a lot of noise lately about wanting to work with the open source community, and there were a lot of headlines this summer, I think it was, when Matthew Szulik met with Steve Ballmer or it was reported that he had. What do you think about this, I mean, would you be if you were in his position, would you be willing to sit down with Steve Ballmer and talk about working with Microsoft?

BY: You talk to everyone. As my father points out, talk is cheap. But the key way to think of it is unless you sit down with everyone, you will not be properly educated. So to, you know, arbitrarily you ideologically refuse to talk to someone, doesn't do anyone any good. It keeps you uninformed or uneducated and it keeps your -- whoever you might sit down to talk to uneducated. So you know, there may be opportunities for Microsoft and the open source world that benefits Microsoft, that benefits -- and it would benefit the open source community and of course, would benefit consumers of software.

The operative word there is that there might be, I don't know if there is. I do know and I would encourage it -- by the way, can't speak to rumors so I'm not commenting on whether or not Matthew actually spoke to Ballmer, but it would shock me if Red Hat weren't, when they go to conferences, sitting down with IBM, and Oracle, and Microsoft and whoever else was willing to talk to them. But Microsoft are a really, really smart organization. They own a couple of very, very valuable monopolies in our industry and anyone who talks to Microsoft had better count their fingers after they (laughs) shake hands goodbye.

NF: I was gonna ask the next question, I mean, would you trust them if you sat down, given the history of their behavior?

BY: No. It's actually, fundamentally no and no one at Red Hat is particularly naive on the topic of Microsoft or any other proprietary software company for that matter. Microsoft has a damaged brand in the technology industry, you know. A healthy brand or valuable brand is one that, you know, consumers prefer because it stands for something good, you know.

So we all love Southwest Airlines. You know, we may not like the fact that Southwest won't assign us a seat, but you know, clearly they've done a great job at reducing airline fares across the country -- just as one example of a company that has a strong, healthy brand. Microsoft's brand, at least in the technology industry is a negative brand. People if they get a choice, will choose to use something other than Microsoft because of this phenomenon that people don't trust Microsoft at this point. And that's part of why Microsoft is, you know, trying to be nice to the open source community is they understand this problem and that their customers are voting with their feet whenever they get the opportunity. They are choosing non-Microsoft solutions simply because, you know, Microsoft is -- I don't want to say this too strongly because I'm not an expert on all things Microsoft but, you know, I do worry about them from a brand-value point of view.

NF: The other thing is, I've noticed that as of last week, OSI seems to be working with Microsoft pretty closely and possibly accepting their new Shared Source licenses to the extent of OSI looks like they're trying to change their tone of their Website and so forth by removing the Halloween documents and so forth. Do you have any thoughts on that?

BY: Again, it's a little bit in the same process. OSI would not be a credible organization if they closed their minds to new ideas. And so, if Microsoft truly does adopt a genuinely open source set of licenses, then OSI should not pretend that they are not open source licenses. Having said that, you know, like everything else, your job -- and this is what I love the media for -- the media's job is to be skeptical. It's to review whatever agreements come out and help readers understand the, you know, whether this was on the up and up or whether somehow Microsoft's money has caused people to act differently than they would had Microsoft not had the money that they do to promote their initiatives. Having said that, I have a lot of confidence in Eric Raymond and the rest of the OSI team. So it would shock me if they ever approved anything from Microsoft that wasn't on the up and up.

NF: Speaking of money, one of the complaints I hear a lot about the offerings from Red Hat is the rather large licensing or, not licensing fees but, you know, support fees. Any thoughts how Red Hat could do that differently and still turn a profit or do you think they're approaching it the right way?

BY: I think they're approaching it the right way. The bulk of Red Hat's customers are -- want more and more sophistication from the solutions that Red Hat are offering. Did I say Microsoft? Anyways, (laughs) Freudian slip. Anyway, Red Hat's offerings have to get better. They are not nearly good enough yet. And the people who are demanding these better offerings are major Fortune 1000 corporations around the world. So for the startups, or the Lulus in 2002, our Red Hat made Fedora available so all of the code Red Hat distributes, they distribute under an open source license and as a consumer of open source and Linux technology, that's all I would ask them to do, how they then come up with a business model that enables them to pay for the engineering that major Fortune 1000 corporation need is a separate issue, and one that's between Red Hat and its customers.

So, you know, and it's a market competition. If Novell does better job of engineering better solutions for Fortune 1000 customers than Red Hat does, then those customers will go to Novell and conversely, if Red Hat does a better job, those customers will remain loyal to Red Hat. But building technology, building the next generation of technology is not an inexpensive project. And so, Red Hat's job is to be a player in an open market where they don't control the customer, they can simply offer services and convince the customer to use their services rather than someone else's.

More on page 4...

NF: Okay. Just out of curiosity, what are you running on your desktop these days?

BY: On my desktop, I run Apple Macintosh OS X. I could not bring myself to go to Windows, of course. And one of the little known facts -- 'cause, you know, I'm a skinny guy who wears glasses, everyone assumes that I'm an engineer by training -- I'm an old typewriter salesman, by training. So maintaining my own Linux desktop was just something that I didn't have the training to do. So I actually went and sat down with a bunch of the guys at Red Hat; this was back in 2002 and asked them what I should run, given that I was not going to, you know, I didn't have the Red Hat engineering team to support me anymore. And their answer was that I should go to the Mac OS X BSD-based operating system because it would give me most of the advantages of Linux without imposing the need to be my own sys admin.

Having said that, that's just my case. Of course, all the engineers at Lulu are all using Linux-based desktops, and Linux is hugely powerful and Red Hat, of course, is hugely successful on the server side. And that's of course where our use of open source and Red Hat has primarily been focused.

NF: What about, so what are your feelings on Linux desktop these days? I guess, you probably don't spend a lot of time evaluating it personally?

BY: No, but what's interesting is a view I held and now that I'm not a director of Red Hat, I suppose I'm allowed to say this again. And unless Red Hat disagrees with me then I don't want to speak for Red Hat. This is a Bob Young opinion and not a Red Hat opinion.

NF: Right.

BY: But it is that -- do you remember way back when Scott McNealy and Larry Ellison -- this is '95 -- announced the network appliance and said that intelligence would remain on the server and that we were all going to start using appliances to access all of the content and the functionality on the server and we're gonna get out of this mode where you had to have a system administrator for every 20 operating system users? Or you had to turn your users of PCs into system administrators themselves, because, you know, these were complicated pieces of technology and it was easy to break something, to do something wrong in it, and so you had to have a whole team of sysadmins.

McNealy and Ellison's vision was to make an appliance that wasn't adaptable; that make the server the adaptable piece in it. And suddenly, the casual user of some piece of computing technology wouldn't have to become a computer expert in order to use the technology. So if you're a lawyer, you should be an expert in the law, you shouldn't necessarily have to become an expert in Windows, you know, desktop technology just to do your legal work. And that was the vision of network appliance.

I'm convinced that that is still, and we're beginning to see it. The future of Linux on the desktop which is, and this is actually interesting enough, it's historically true, as well. Operating systems typically don't displace existing technology if they become popular, it's because they become the popular solution on the new technology model. So, you know, Digital Equipment Companies, VMS operating system did not become popular on IBM mainframe computers. It became popular because Digital convinced the world to move towards minicomputers, and it became the most popular operating system on the minicomputers that Digital was responsible for. And in the same way, Windows, you know -- Microsoft did not convince the world to unplug VMS and put, you know, sort of a Mac OS -- not Mac OS. MS-DOS, MS-DOS on Digital minicomputers.

They convinced the world to put MS-DOS on this new technology model, being the PC. It was again, for Linux, to try and convince the world to unplug MS-DOS and Windows off of their perfectly good PCs and put Linux on it instead, that's a hard thing to do. But if the network appliance vision of the future comes true, then it makes perfect sense too, that open source operating systems such as Red Hat, or Slackware -- whatever -- would become the dominant operating system on that kind of appliance. And that's actually what we're seeing. Best single example is the Tivo box on your desktop, on your set top, on top of your TV which is a network appliance. As in it's an appliance, it's not something you asked to program or, you know, install software on; it comes as an appliance. It does its thing. And of course, it runs Linux as its operating system.

NF: Do you think that the home PC is a phenomenon that's going to continue for the next, you know, say 10 years out?

BY: Yes. In the same way, though, when you think about that people are still selling mainframe computers, so, but it will become a smaller percentage of the total number of computers we carry around with us. So, in the late '80s or through the '90s, PCs were where the bulk of the dollars went to when you were buying computer functionality. Today, the bulk of the dollars are going to things like the Treo that I've got in my pocket. Things like digital cameras that I can plug into my PC. Things like my car. You know the map tool that I get in my rental cars, you know, that tell me when I'm getting lost in a new city or how to stop getting lost. These are all computers that are not PCs, they are all network appliance like computers, and that's where the bulk of technology dollars are going today.

NF: You mentioned earlier in the interview that you looked around for things to spot through an open source. So, are there any areas where you see things missing in open source or where there needs to be more development focused?

BY: How do you answer that?

NF: I mean, personally, I can think of a few things I'd like to see improve but I'm curious where you're...

BY: Yeah. The broad answer is there can never be enough. So it doesn't matter how many tens or hundreds of millions of dollars are going into building better open source applications. It would be nice to see even more dollars going into building open source applications. And the challenge we're up against is the proprietary software industry is so big and so lucrative; a business model that, you know, we're playing a fair amount of catch-up. Microsoft have, however, many tens of billions of dollars to dump into their R&D effort.

We, on the other hand, have equally large numbers of dollars but not as easily focused. So the bulk of the open source engineering is coming from the engineers using the code. So it fits. The bulk of the engineering that's being done at GE or at Southwest Bell or at Boeing; building tools that they and similar users across the Internet need to use. The dollars that IBM and Red Hat and Novell are putting into engineering open source projects or MySQL or you know, these kind of guys is actually the smaller part of all the resources that are being put into building open source tools.

So, the short answer is, yeah. But to try and give you a specific example, I'd sort of duck on that one. I'm not really sure. I can think of a hundred examples of things that could use more investment. I can't think of -- I wouldn't want to prioritize, so one of my favorites is some of the tools that Lulu are using. We would love to see Postgres, get further investment and get some better tools, for example, commerce shopping cart tool built in an open source model so that we didn't have to build our own commerce and shopping cart tools at Lulu. In order to have open source commerce and shopping cart, you know, there's a million and one proprietary shopping cart tools we could buy. There aren't that many very good open source shopping cart tools we could buy.

NF: Are there any new projects that have come out of Lulu that you're sponsoring that we could find like on, you know, SourceForge or anything like that?

BY: Yes, is the short answer. But I'm not gonna speak to them because I'd probably make a hash of the story. Keep in mind that I'm the typewriter salesman in this thing.... Just as an example, there's a project called, that was called Tiki pro, I think it's just changed its name. And it's a content management tool that we use quite actively, and that we are stressing; as in we're looking for functionality beyond what the original project was offering. And so we've been investing in that project, in some of the project leaders across the Internet to try and move that project forward faster so that we could get the benefits out of some of these content management features that would save us again from having either to build our own or to consider proprietary solutions.

NF: Back earlier this summer again. Eric Raymond made a comment that got reported pretty widely, which was that we don't need the GPL anymore. Any comments on that idea?

BY: Yeah. I couldn't disagree more. And again, keep in mind, I'm not picking a fight with Eric 'cause I have a huge amount of respect for Eric. He is one of the most valuable members of the community; and unlike a lot of the members of the open source community who get quite ideological about things, and when they disagree with another member of the community, they tend to get emotionally caught up in their disagreements. I'm more than happy to disagree in a very friendly, supportive, you know -- this is my friend and I'm allowed to disagree with him. Wait, so don't present this as anything other than me disagreeing with an interpretation of a friend of mine and it is simply that this, and Eric and I are approaching this from different perspectives.

Eric's looking at it from a legal, implementable point of view. And there is no question that the GPL is a badly structured piece, uh, legal document. And it's badly structured because it's not just a legal document, it's also a statement of ideology. But I'm a sales guy, so I don't actually care about the details of things. I care about their impact. And what makes the GPL so valuable is not how it's written or what it legally enforces, it's the fact that it's a brand.

The moment you say your software project is GPL, everyone knows exactly what that means and they don't have to read your license to understand whether you're sneaking in the back door, or whether you have some sort of ulterior motive to what you're doing. If it's GPL-ed, it is GPL-ed and we all know what GPL means. So I would strongly argue just the reverse: that we will never outgrow the GPL. And that the GPL will continue to be the dominant open source licensing model, not because the GPL is the best license, but because it is the standard license.

So to argue against the GPL is like trying to argue against MS-DOS. If by saying we need a better operating system, arguing against MS-DOS in the late '80s, early '90s, because everyone knew we needed a better operating system than MS-DOS for the PC. But everyone also knew that MS-DOS was the standard. If you built your software to run on MS-DOS, it would run on hundreds of millions of personal computers around the world. If you built your software to run on "a better operating system" than MS-DOS, say the BeOS, you were addressing a market that might have had a few thousand installed computers.

Well, the same principle with GPL. You write, you license your open source project as a GPL project, everyone knows what it means. Anyone who might want to contribute can contribute without reservation, without having to get a lawyer out to read the license and to try and understand why the comma in the 13th paragraph, whether that comma had significance or not. The GPL, we all know what the failings of the thing are. But we also know that it's the standard.

NF: What do you think of the efforts to rewrite the GPL? Any thoughts on the coming version, GPL 3?

BY: Not really. Because I'm not a lawyer and I have not been involved in the project. It's both good news and bad news. If they do it well, it'll be a great thing because it'll address some of the weaknesses in the GPL. If they do it badly, they could screw up a really good thing. So, on the one hand, I'm nervous; on the other hand, I'm optimistic. But I have to admit, I haven't been staying up with it at all.

NF: Anything else in the tech industry, either open source or otherwise you find really exciting these days?

BY: Yeah. The thing I find exciting is the Internet itself. As an entrepreneur, I look at the Internet as being the Wild West. And so a lot of people are starting to look at the Internet as being this established, you know, sort of a stable environment that, you know, that all the big opportunities, you know, that were out there, were out there in the 1990s and now it's time to get a job working for Yahoo! or Google if you want to work on the Internet because there's no real opportunity for the little guy anymore. And my take on it is just the reverse. It's that the big guys on the Internet today are gonna look small compared to the big guys on the Internet 10 years from now. And that they will be named, they'll be household names on the Internet 10 years from now that you and I have yet to hear of today.

It looks a little bit like the PC industry back in the late 1980s when the big names were people like Ashton-Tate and WordPerfect and, you know, the list goes on and on -- 3Com. And today you're going, "Whatever happened to those guys?" And meanwhile, guys who were trivial players in the mid '80s -- Dell were only founded in something like 1985 -- are now the dominant players. And yet, in 1985-86, everyone in the industry thought of it as a very mature industry. I see the Internet being where the PC was in 1986. The other thing is open source. There was a statistic -- I'm sure you saw it; I think it was out of Gartner -- talking about how 56% of all software programmers were working in open source tool these days. Did you see that number?

NF: Yeah.

BY: Single, most significant statistic I've seen in open source in a long time because where the engineers are is where the innovations come from. And Microsoft worked like dogs in the 80s to get all the software programmers to program for Windows. And if all the programmers are starting to program in and around open source, then it's only a matter of time before open source tools continue to get better and continue to get more widely deployed. It's no longer a question of 'if,' it's simply a question of 'when.' Because once the programmers have voted with their feet, you know the rest of it just kind of happens almost automatically.

NF: The lure of open code is just too great to ignore.

BY: That's right! It improves your ability to build better tools because you're not guessing at what's going on inside the database or inside the Web server. You're actually looking at the code and when you build your application on top of it, you're not guessing at whether you're doing the right thing. You can actually watch the code cycle through and you can go, "Jeez, now I see why it hiccupped on line 56." You're not guessing at it anymore and that's simply a dramatically better way of building tools, especially building Internet-based tools than trying to do it on proprietary software where you actually don't know what's going on.

NF: I guess it brings us back to why Microsoft is trying to figure out just how much they can open without giving up their licensing.

BY: That's right, that's right. It's very encouraging, I guess, in a nice way to see Microsoft on the defense. I picked up a newspaper in Toronto the other day and a little ad on the top, the masthead of the business section, I think it was. I think they called them, "earlug" advertisements; so you have, you know, things like "Business Section" across the top and little ads on each side of the business section. And this little ad was "Windows vs. Linux: Get the facts -- www.Microsoft.com/getthefacts" or something. And I'm going, "Wow." Can you imagine 10 years ago Microsoft even admitting that Linux existed and now they are buying ad space in conventional media outlets like daily newspapers in order to try and fight Linux? It's Eric Raymond's line, I think he stole it from Gandhi, which is, "First, they laugh at you; then they fight you; then you win," or something. Clearly, in stage number 2 of that progression.

NF: Do you have any other thoughts or anything you'd like to say I haven't brought up?

BY: No, Joe. I really appreciate your interest in the story. The key theme to my leaving Red Hat had nothing to do with Red Hat and everything to do with Lulu. On the one hand, of course, my own personal reasons to get out of bed in the morning, which is I'm an entrepreneur by, not just by training but by instinct. And Lulu's -- it's a lot more fun to get out of the bed in the morning these days, focusing on Lulu, than it is going to play golf or, for that matter, going to Red Hat board meetings even though I love Red Hat and I take a huge amount of pride in what those guys are achieving.

Category:

  • Business