TCS Daily

Why Open Source May Be Doomed

By Megan McArdle - October 6, 2003 12:00 AM

I have to admit that I was never much of a believer in open source. Maybe my business school coursework rendered me blind to the glorious vision of a "gift culture" in which people contribute their work to a decentralized development project like Linux for honor instead of money. Or possibly I'm just too thick to understand how cutting off a multi-billion dollar revenue stream from software sales, without putting anything else in its place, could be good for the software business. Whatever the problem, I never quite believed in the fairy tale world they promised in which we'd all get an operating system that was better than Windows in every way, for absolutely no money -- not even when IBM started retailing Linux PC's and the juggernaut of fabulous free operating systems seemed unstoppable. But I confess that in all my skeptical musings, I did not imagine that Linux might be brought down by something even more prosaic than a lack of funds: a lawsuit. 


Yet that's looking ever more likely. SCO, which makes a proprietary version of Unix that Linux competes with, has filed a suit against the manufacturers of Linux boxes for copyright infringement. IBM, which has been promoting Linux relentlessly, is now announcing a countersuit, but it centers mostly on side issues, rather than the key question: did one of Linux's thousands of volunteer developers illegally stick code stolen from SCO into Linux? Though those who have seen the code in contention say that SCO probably has a case, it doesn't seem to be much of a case: the stolen bits seem to be fairly trivial and easily replaced. But of course, the object of this lawsuit is not to stop Linux from using the code; it's to stop Linux from eating SCO Unix's lunch. And it seems to me that it's very likely to succeed.


Don't get me wrong: for all my skepticism about open source as a business model, I think Linux is a great operating system. The decentralized development model, in which volunteers are constantly building, refining, and debugging the code has produced an unquestionably fine product. And, of course, it's free. Nonetheless, if SCO wins this lawsuit -- or loses on technical grounds other than copyright infringement -- I think you'll see corporations taking a pass on open source software. Because for corporations, the real problem with this lawsuit is not a few lines of stolen computer code, which is why HP's attempt to stop the damage by indemnifying its Linux customers against SCO is unlikely to work. The real problem is this: if you're an IT manager deciding whether or not to purchase a Linux machine, how can you be sure that those stolen lines are the only ones?


Corporations simply can't afford the risk of a lawsuit, even if the cost of a non-open-source OS is several hundred dollars higher. At the corporate level, lawsuits are expensive and distracting, even if you win. And at the IT manager level, telling the board that your hot new installation just embroiled the company in a legal battle is a career killer.


For Linux, that's bad, bad news. The outcome of the Windows/Mac showdown seems to indicate that the company that owns the corporate desktop owns the marketplace. So the last vestiges of those glorious internet bubble dreams of an economy based on free stuff may finally be at an end.


But there is no cloud without a silver lining, and contemplating the possibility that Linux will fail to take over the desktop offers some valuable lessons for those of us who spend our time trying to figure out what the world is going to look like in five or ten years. The first is that you should never predict the imminent demise of the old way of doing things without examining what benefits the old way might offer that the new way can't. And the second is that as long as inhabitants of the virtual world have to occupy physical space, the offline world is going to exert considerable power over those virtual interactions.


For example, consider the auto dealership. When the internet began to take off as a retail platform, I -- and many other people smarter than I, whose job it was to analyze the market -- was confident that the auto dealer was on his way out. After all, the value proposition that most car salesmen offer their customers is . . . ahem . . . rather limited; it consists of concealing from the purchaser the true cost of the car and the terms of the financing agreement in order to squeeze every last possible dollar out of them. Wouldn't it be much more efficient, I thought, just to test drive the car at a company-owned showroom, and then order the one you want online?


Well, yes, for the customer. But it turns out that auto dealers offer some very valuable services to the car manufacturers. Because I was totally ignoring that side of the equation, I didn't see that dealers were in fact occupying a crucial niche.


Auto manufacturers, it emerges, are not (yet) able to custom manufacture cars on demand, for a number of reasons. They have to produce cars based on their best estimate of future demand. They thus require a distribution system that is highly plugged into the local market, and can vary prices and terms to best move the cars that have already been produced off the lot. If they were selling cars on the web, auto manufacturers would have to set one national price, which means either forgoing significant revenue from affluent areas by pricing to the nation's poorest region, or lose sales in poorer regions by pricing to the upmarket.


But of course, the value proposition they offer the auto companies isn't the only reason that dealerships survive. They also survive because in their generations of operation, they have built up significant power -- both in negotiating with the auto companies, and in lobbying the government. That's why you can shop online all you want, but you still have to get your car from an auto dealer: dealer trade associations lobbied state governments to pass laws making it illegal to do otherwise, and bullied the car companies into going along. As long as GM has to have a physical headquarters, and its cars have to ride on physical roads, it can't ignore the law just because the transaction is taking place in the brave new virtual world.


So with Linux. Those who saw open source as the inevitable revolution concentrated only on the value propositions that Linux could offer on price and performance. They weren't paying as much attention to other considerations, such as indemnity, that might be even more important than performance to the marketplace, because after all, what open source developers are good at (and interested in) is technological revolution, not liability planning and insurance premiums.


They also tended to assume -- as did I -- that decentralized or virtual structures were somehow immune to legal action, even though that assumption has proven incorrect time and time again over the last five years. While many have focused on closing off the last avenue of attack -- building peer-to-peer file sharing networks like KaZaa, for example, after Napster's centralized server-based system was shut down -- fewer have come forward to say "We'll have to operate in the same legal environment as offline businesses. Let's tackle that." But fighting a rearguard action against the last incursion of the offline world is a losing battle in the long run. This lawsuit is bad news for one technology, but a good opportunity for us to finally learn a lesson that we seem to be having a lot of trouble absorbing: if you want to overthrow the current regime, it is best if you first learn exactly what you're dealing with.

TCS Daily Archives