LIKE COUNTLESS OTHER urban, Web-savvy media consumers too lazy to leave their apartments, I greeted last month's demise of Kozmo.com with something alarmingly close to actual grief. Not because it suggested more dark clouds ahead for the already gloomy dot-com sector, nor because I had some kind of direct financial stake in the online delivery service itself. I found myself mourning for Kozmo because I had become an avid user of the service, and had begun to wrap a subsection of my daily impulse buying habits around its bright orange plastic bags. Just as Amazon had transformed my book-purchasing habits two years ago, Kozmo had become my default source for DVDs, games, and late-night snacks -- almost entirely eliminating the longstanding relationships I'd had with Blockbuster, assorted software stores, and the twenty-four-hour deli on my corner. Waking up one day to find that the site had vanished into thin air meant that -- for the first time -- the rhythms and rituals of my consumer life had been interrupted by the dot-com bubble bursting. The collapse of Pseudo, Pets.com, Boo -- not to mention the thousand failed companies I'd never heard of -- were really only passing headlines for me. But Kozmo had entered my life in some small but significant way, and when they shut their doors, it left a little hole in my daily routine that I've had a hard time filling.
I was reminded of that hole when I read the transcript of Craig Mundie's much remarked upon speech from early May, in which the Microsoft executive laid out the company's "shared source" software development model, and took a few passing swipes at the open source movement -- and particularly the GNU General Public License (or GPL) that governs many open source projects. The antagonism between Redmond and the open source community is heated enough that anytime Microsoft so much as mentions the word Linux or Apache in a public statement, it triggers a deluge of close readings, counterattacks, and thousand-post threads on Slashdot. But Mundie's comments were particularly interesting because they also attempted to put an interpretive frame around the dot-com implosion -- a frame that explicitly linked the shortsightedness of the bubble economy to the perceived liabilities of open source models. "In this sense," Mundie argued near the end of his talk, "open source software based on the GPL mirrors the dot-com business models that proved the least successful during the past year. They ask software developers to give away for free the very thing they create that is of greatest value in the hope that somehow they’ll make money selling something else. In effect, it puts at risk the continued vitality of the independent software sector."
Those are fighting words, and they outline a smart position for Microsoft to take -- but they're also based on a delicate chain of associations. The "least successful" dot-com companies were not necessarily involved directly in the open source software movement. (For all we know, the folks at DEN and Pseudo were serving their free content off of NT boxes.) What those dot-bombs shared with open source software was a willingness to give away for little or no cost products that a company like Microsoft might be inclined to charge for. Kozmo, for instance, was famous for charging so little for its goods that their revenues didn't even cover the salaries of the delivery guys. It's this tendency to give things away for free that grants Mundie his parallel between the two sectors, letting him write the posthumous history of one field as an indirect way of analyzing the prospects of another. His autopsy of the bodies in the dot-com graveyard sets the stage for a prediction about the future health of open source.
Not surprisingly, his prognosis isn't good.

ARE THE FAILED business models of Kozmo or Pseudo truly analogous to those of Red Hat or VALinux? Not necessarily. For one, most software companies in the open source world do charge a clearly defined, upfront fee to their customers. Red Hat isn't exactly handing out its shrink-wrapped flavors of Linux on street corners, after all. They're charging forty dollars for the package, even though they took on almost none of the overhead involved in creating an industrial-strength OS. Technically, they're charging for distribution, packaging, and support, not the actual software, but the end result is someone buying an OS from a vendor who didn't pay the programmers who created it. In a funny way, you can think of the commercial open source world doing precisely the opposite of the revenue "strategy" that Mundie attributes to the dot-coms. They ask you to pay for something that they themselves didn't pay anything to create.
Then there's the thornier question of the GPL. It is certainly true that the GPL presents serious challenges to anyone who wants to "rent" software to another party, à la Microsoft, without handing over the ownership rights in the transaction. "The GPL mandates that any software that incorporates source code already licensed under the GPL will itself become subject to the GPL," Mundie explains. "When the resulting software product is distributed, its creator must make the entire source code base freely available to everyone, at no additional charge. This viral aspect of the GPL poses a threat to the intellectual property of any organization making use of it." "Viral," of course, is the expected meme here, but the analogy might be better drawn to modern genetics -- putting a given piece of code under the GPL is like tinkering with the stem cells of an organism: The choice you make gets passed down to all the descendants, no matter how the environment changes from generation to generation. For software developers, this has significant implications. Let's say you build the digital equivalent of a better mousetrap -- a grammar checker that actually works -- and then decide that it's best served by the GPL. With that single decision, you've effectively eliminated the possibility that any of your program's descendants will get integrated into Microsoft Office at any point in the foreseeable future, assuming the foreseeable future doesn't include a forced collectivization of the entire Redmond campus and all of its intellectual property.
Now consider the implications of the GPL from a consumer perspective. What should attract an end-user to a piece of GPL'd software? Here we get to the heart of Mundie's critique, and the heart of its Microsoft-centrism. The grammar checker that doesn't get integrated into MS Office is only an evolutionary dead end if it also fails to survive and evolve on its own. Given that only a finite number of software tools will eventually make their way into the MS toolbox, any software buyer interested in long-term viability has to look very seriously at tools that belong to the open source community, and not a proprietary vendor -- because open standards tend to have a longer life than proprietary ones. Eric Raymond refers to this attribute of the open source model as "risk-spreading": By widening the pool of people who are free to develop the code base, you narrow the chances of the code's disappearing because a single company stopped supporting it. In his essay "The Magic Cauldron," Raymond describes "a model in which open source functions not so much to lower costs as to spread risk. All parties find that the openness of the source, and the presence of a collaborative community funded by multiple independent revenue streams, provides a fail-safe that is itself economically valuable -- sufficiently valuable to drive funding for it." The companies that package the code base may come and go, but the software lives on.
Once again, the course followed by the dot-com bubble companies turn out to be the exact opposite of the one pursued by the open source world. You can think of Kozmo as a kind of proprietary data delivery platform. (The platform is partially built out of guys on bicycles, of course, which is why it was always a little weird to think of Kozmo as a pure Internet company in the first place, but that's another story.) When I decided to build a consumer relationship with Kozmo -- giving them my credit card, letting them cookie my browser, building a purchase history with them -- that commitment was predicated on their future ability to stay solvent, just as it would be if I took the time to enter my medical history on WebMD or my golf scores on Golfweb. When Kozmo failed, the service that I had grown so accustomed to disappeared overnight, the same way my data disappears if Golfweb or WebMD shuts down. While most dot-com pricing was certainly competitive -- nothing's more competitive than free -- the consumer proposition was a risk-narrowing model, rather than a risk-spreading one. You build a complicated and long-term relationship with a single provider, and hope they can keep the bank balance flush. If they can't, tough luck.
This is the ultimate flaw in Mundie's argument. If creative destruction is a hallmark of the new new economy, if we're fated to live through increasingly turbulent economic times, then open source software isn't one of the causes behind that volatility. It's an antidote.

 

THERE'S ANOTHER POTENTIAL antidote, of course -- you can always choose to buy from a giant, semi-monopolistic, closed source software company like Microsoft, and feel confident that your code will continue to be supported for decades. Obviously, this is a model that has worked extremely well for Mundie and his colleagues, and so it's only natural that he should encourage the rest of the software world to follow it. The problem is there's a finite amount of room for giant, semi-monopolistic, closed source software companies -- in fact, there may only be room for one, and Microsoft doesn't look like it's going anywhere soon. What the commercial open source sector has decided is that it's worth it to trade the high margins of proprietary code for the dramatically reduced development and marketing costs of GPL'd software. You make your money by being an expert packager, and let the entire development community do the heavy lifting of actually building the code base. (Raymond calls it "Give Away the Recipe, Open a Restaurant.")
Will any of those software restauranteurs become the next Microsoft? Probably not. But they could very well become sustainable business that turn a tidy profit, and in the process help create a more open and diverse software ecosystem. And that's the final irony of Mundie's analogy between open source and the bubble economy: The insane late-nineties valuations of the dot-coms were based partially on the hope that some of those companies were in fact "the next Microsoft" -- a juggernaut in embryo, a company that might some day sport a $100 billion market cap, and thus be a bargain at $10 billion. (No doubt some of the Linux companies benefited from those hopes as well.) The crash happened because a lot of investors woke up one day and realized that they were holding shares in the next Pets.com, not the next Microsoft. Most of those companies have dealt with that cruel realization by going out of business. But the open source companies have another option, one that is true to the values of the movement, and in synch with these more modest times. They can position themselves as the anti-Microsoft, not in the sense of another proprietary wannabe world conquerer, but in the sense of presenting a challenge to the entire model that Microsoft's success is based on -- including the company's epic profit margins.
"Less killings and more livings" is sometimes thrown around as a motto for the post-Napster music industry: more working musicians paying the bills with their art, but fewer mega-millionaire superstars and record industry execs. The same slogan applies to the ongoing experiment of commercial open source, an experiment that seems to be plowing ahead full-steam despite the NASDAQ correction, despite the doomsayers on Wall Street -- and despite Microsoft's misguided attempts to link the open source movement to the ghosts of the long boom.

Steven Johnson is FEED's editor-in-chief and the author of Interface Culture: How New Technology Transforms the Way We Create and Communicate.

Other articles by Steven Johnson

Back to Article