An Interview With Eric S. Raymond

Eric S. Raymond's The Cathedral And The Bazaar paper galvanized the open source community in 1997 while it simultaneously turned the mainstream computer industry's structure on its ear. Raymond argued effectively that a loose organization of talented developers bound by a common goal (referred to as "the bazaar") could produce higher-quality software than well-funded organizations could in their "cathedral". As if to prove his thoughts correct, the open source community has responded in recent years by producing powerful software products used across every country and every vertical category. These products include the Apache Web Server (60+% market share), the Linux operating system (currently over 30% Internet server market share), and widely used programming languages such as Perl, Python, and PHP.

From our vantage point, it seems but a matter of time before the open source community focuses its efforts upon mobile devices and applications, so we thought it would make sense to speak to one of the community's leaders. Never one to shy away from a thoughtful discussion, Mr. Raymond was kind enough to take time out to discuss the open source software movement and his thoughts on the future of mobile computing.

WDN: For starters, what are your thoughts on emerging mobile technologies such as WAP, location-based services, and Bluetooth? Are you excited about the possibilities? Any reservations or doubts as to their potential?

I'm excited about the potential of wireless broadband, of constantly-on Internet access that you can carry with you. So I think Bluetooth is very cool, and I'd be friendly to location-based services based on open stasndards.

On the other hand, I'm not thrilled by what I hear about WAP. It's too limited, and it's essentially a vendor-controlled proprietary gloss on HTTP. No thanks. Just give me TCP/IP and let me decide what I do with it.

WDN: Do you use any of these technologies yourself (i.e. Palm, WAP phone, etc.)?

A friend gave me a Palm III. I don't use it; I travel with a Sony VAIO instead. Basically, if I can't write code and handle email with a device, I'm not interested in carrying it around.

WDN: Are you still coding? If so, what tools or products have impressed you most recently?

I am definitely still coding. I'm enjoying my VAIO 505TR a lot. I'm doing a lot of hacking in Python (the open-source scripting language for when you outgrow Perl). And I've been a big fan of the Logitech line of optical trackballs; they have true elegance.

WDN: While most server admins and Web developers are very familiar with open source projects by now, talk of this type of development is new to the wireless industry. Would you mind briefly explaining the advantages to this development methodology and how commercial companies can develop revenue streams from open source products?

The biggest advantage of open-source development is the peer-review effect. You get super-reliable code. You also get customer trust because they know you're following good practice -- and that they can't be locked into bad choices by closed code.

Commercial companies can reap huge market wins from open-source code in several ways. One is by being a supplier that builds infrastructure around open source and open standards. You'll clobber your proprietary competition, in the same way that and for the same reasons that TCP/IP clobbered proprietary wide-area networking in the early 1990s -- and the same way Linux is clobbering Microsoft in servers and Internet appliances now.

Another is by hooking up to an existing open-source infrastructure and selling services and content over it. Anybody who does business over the Web is executing this model. Over 60% of the Web is served by Apache, the open-source webserver. E-commerce is getting huge exactly because nobody (not even Microsoft) can put proprietary roadblocks on the information superhighway.

WDN: What significant issues do you see that still stand in the way of open source projects?

Submarine patents held by hostile closed-source vendors could turn into a nasty problem. Other than that it's basically all a question of perception -- corporate America is still getting to grips with the idea that open source is an economic win. There's still a bad tendency out there to overprotect supposedly valuable intellectual property that is really just a ball and chain dragging on your potential growth -- to believe you're in the secrets business when you're actually in the snarts and service business.

WDN: PDA/Handheld operating systems are very proprietary in nature and offer unfamiliar, closed APIs to developers. From my vantage point, this space seems to offer open source developers a unique opportunity for two reasons. First, they won't have to compete against an entrenched operating system. Second, a modular OS like Linux could offer a set of common APIs across pltforms (from PDA to desktop to mainframe). Are open source companies interested in aggressively pursuing this space?

I think if you were a PDA manufacturer that Red Hat or Corel or Caldera would be more than happy to help you port Linux and then maybe co-brand or Linux-brand the device. This kind of thing could be a big win-win-win for developers, customers, and vendors alike. I expect to see more of it as soon as PDA vendors figure out that trying to sleep with Microsoft is suicidal and being in the OS business yourself is just stupid.

WDN: Have you seen the Transmeta technologies in action? Some have said the secrecy surrounding the project was simply great marketing. Others contend that there's nothing really new going on that hasn't been done before. Since they will be providing support for Mobile Linux (and perhaps other open source products), what are your thoughts on their technologies and their chances for success?

I haven't seen their stuff in action. I'm excited by the potential, but a little disturbed by their refusal to publish their benchmark source and protocols for third-party evaluation. That makes me wonder if they can back up their brag. On the other hand, it's not smart to underestimate the caliber of talent they have there (and I don't only mean Linus Torvalds, either).

I'm hoping it will all come good in the second-generation chips. At this point my dream laptop would wrap the VAIO's case design and form factor around a fast Crusoe, running Linux of course. That combination would be nearly as good as sex :-).

WDN: What are the leading opportunities that you foresee for open source projects (in terms of potential applications or markets as yet untapped)?

The biggest untapped space, of course, is the desktop; we're only at 1-4% there. I say that not because I think that's the most interesting place to be, but because open source is well on its way to owning everything else. The trend curve on our adoption in servers is surprising even to me; IDG just reported a Linux growth rate of 166% over the last year up to over 25% (34% in Internet servers). On that trend the closed Unixes are history and Microsoft is looking down the twin barrels of doom in early 2001 or so.

In the areas most likely to be of interest to your readers (PDAs, appliances, wireless devices) Linux and other open-source projects are bound to win big on simple economics. Hardware price points are so low that closed-source platform licenses price themselves right out of the market (Microsoft has found this out the hard way with WinCE).

On the software level, I think we'll see a continuing push into applications and end-user interfaces. The formation of Eazel is a pretty good harbinger -- put the guys who did the Mac interface together with Linux and $11 million in funding and you will get something interesting.

WDN: Alot of people look up to you and the work you've done...accidental revolutionary and all that :-) Who do you admire in the technology/business world today?

Well, Linus Torvalds of course. Becoming his friend and co-conspirator and minister of propaganda hasn't undermined my respect for his amazing talent and audacity and people skills. I still admire Richard Stallman and try to maintain our long friendship even though I think he's out to lunch on some key issues. Larry Augustin (founder CEO of VA Linux Systems; I'm on his board) continues to impress me by combining business savvy with a hacker's pure vision of a better, open-source-centered computer industry. I admire Tim O'Reilly's near-magical ability to turn technical publishing into an expressive art, to do books about software and the Internet that are works of candor and beauty. He, like Larry, is an entrepreneur with deep integrity and a vision that genuinely aims at a better future.

In the wider world, I'm a fan of Clayton Christensen (author of "The Innovator's Dilemma"). I met Jeff Bezos briefly once and would like to chat with him sometime; if anybody can actually pull off the dot.com thing in a sustainable way I'd bet on him to do it. I'm not sure who the strategic brains behind Cisco are these days, but I admire them too. I'm learning a lot these days from the work of the economist David Friedman.

I continue to admire, as I have for all twenty-four years of my career, the original Unix guys. Ken Thompson; Dennis Ritchie; Brian Kernghan; Doug McIlroy; and all the rest of that original Bell Labs crew. These men created the software tradition that led to the post-1985 Internet and to Linux. Our debt to them is immense, and grows every day.

WDN: Most engineers and technologists are known for their dubious writing skills. Have you always been interested in writing or was it a skill you developed out of necessity?

Always. I write as naturally as I code. I can't imagine not doing both.

Interestingly enough, by the way, the rule that most programmers are clumsy writers breaks at the top. All the hackers I know well who I consider among the mightiest wizards are also remarkably expressive and able writers; RMS, Larry Wall (author of Perl), Guido Van Rossum (author of Python) and Linus himself are all examples of this.

WDN: Care to step out on a limb and paint a picture of the technology world in, say, 2002-2003?

Well, that will be two years after the crash of Microsoft in early 2001 :-). Windows will be a fading legacy system that never made it off aging 32-bit hardware; by 2003 people still running it will be regarded with the same faint pity and amused condescension that people still using dedicated word processors or Commodore-128s meet today.

The hardware market will look like a three-cornered war between 64-bit Intel boxes descended from today's PCs, "gaming" systems with auxilliary keyboards and more pixel-per-second rendering power than today's supercomputers, and super-appliances based on Crusoe and StrongARM chips. The Intel boxes will probably be losing, slowly.

These machines will all have big flatscreens, by the way. Today's bulky boxy monitors will look already look quaint and absurd.

Linux will be everywhere, in thicker or thinner disguises. Turnkey versions will run the appliances (and your cellphone and the web browser on your refrigerator door). Your 64-bit-monster PC will boot with a penguin logo into a desktop you won't easily be able to tell from Microsoft Office (except that it doesn't crash). You probably won't know how to get to the Linux underlayer on the PlayStation VI in your TV room -- but your kids will.

The Internet will be everywhere, too. The Bluetooth network or a descendant will be nearly as ubiquitous as the cellphone net is now; all new laptops and PDAs will come with built-in aerials. Shops, restaurants, and other public places will be starting to install Bluetooth relays as a cheap loss-leader to retain business, piggybacking the "free rider" use on the broadband links they'd run for B2B communications anyway. (One of the public controversies will be whether responsible businesses should block porn downloads over Bluetooth.)

Whether Gush or Bore is elected, politicians talking about the net will continue to be clueless and annoying.