WINNER TAKE SOME

"Winner Take All" coverage of trends in computing is used so often it has become reflexive for most of the press. Most computer reporting today takes its cue from the Mother of All Holy Wars, the question "Which Operating System is the Very Best One?" Covering all software developments as if they were operating systems creates a false sense that one piece of software or the other will inevitably "win" and the others will "lose".

In the case of the browser war, "Winner Take All" reporting obscures the fact that the near-term outcome of this "war" is obvious to anyone with a stake in it - neither Microsoft nor Netscape will win, and the reason has nothing to do with their various strategies or tactics - "winning" the browser war in the near term (the next year, say) is simply not an option.

While the press (and the companies themselves) strive to create the illusion that one of the two of them will eradicate the other, the reality of the situation is this: the open standards the browsers are written to are mostly outside the control of either company - each browsers' performance in handling HTML, email, and Usenet news must conform to specifications their programmers did not invent and which they can only affect but not control. If either company announced that it's browser was going to render HTML tables differently, for example, it would lose users because of perceived incompatibility with accepted norms. Furthermore, where these products do diverge (like Javascript or ActiveX), they produce an incentive for the designers of Web content to ignore the proprietary functions and design Web sites whhich only rely on areas of commonality (like Java). In other words, divergence from the norm can accelerate the use of open standards by the designers who create the content.

I do not mean to suggest that there is not a competition going on here, there obviously is, and its outcome oviously matters deeply to both sides. However, what Netscape and Microsoft are fighting towards is not total victory or utter defeat, but to something much more complex - they are fighting their way to some kind of equilibrium. In the near term, it is likely that each product will develop its core audience, and then fight over the audience in the middle.

Now "equilibrium" is not the kind of word that sells a lot of magazines, but I want to argue that "Winner Take All" reporting has damaged not just coverage of a schoolyard fight over Web browsers - it has damaged a lot of people's assumptions about computers and networks generally, and that the same mistakes about victory and defeat that are characterizing the browser "war" are pervasive in the computer press. This is especially important because browsers are the likely precursors to the development of network computing, thin clients and virtual machines.

Web browsers perform three essential functions of a network computer interface - they standardize interface across multiple hardware and OS platforms, they recognize and interpret platform independant code, and they allow the division of computing tasks among local and remote machines. Because of this, what we are seeing in the browser wars of today is the best predictor of the battles to come over NC, NetPCs, thin clients, and the like. Browsers are a kind of fat "thin client".

So what is the habit of winner-take-all thinking likely to obscure in the coming competition over hardware and software standards for network computers? The main problem with the convenient idea of winners and losers is that the losers rarely go away. (I am speaking here of losing ideas - of course losing companies often go away, but that can have as much to do with marketing as with merit.) According to winner-take-all logic, for example, the mainframe is dead and client-server is everywhere ascendant. Tell it to your bank (or your credit card company, or your travel agent). For mission-critical applications, global networks or 24/7 uptime, the mainframe is still a good solution - mainframes now perform better in their newly restricted domain than they did when they were the "one-size-fits-all" solution. Furthermore, as thin client computing takes off, some of the machines at the center of these thin-client networks are going to look surprisingly like mainframes talking to "just-smarter-than-dumb" terminals.

When losing ideas lose, they often benefit from the struggle, and far from disappearing, they can come back stronger. Two major proposals for seeding the market with NCs revolve around recusitating older hardware: Sun's plan to turn 80486 machines and Hewlett-Packard's plan to turn Xterminals into into Java-enabled NCs. In these cases, progress relies on the leverage provided not by the most current multi-media PC technology but on the very machines those fancy PCs overthrew.

What is true of hardware is true of software as well: simple software, replaced by its bloated descendants, finds new life on ever-smaller devices. My favorite example is Tetris, that age-old game long ago replaced by flashy new PC games, which now exists in versions that run on PDAs. As processing power becomes increasingly ubiquitous, I would not be surprised to hear that Tetris has been ported to a toaster-oven, and despite all the focus on businesses saving money, you can be sure that a thin-client Tetris is in the works alongside downloadable spreadsheet programs.

I would like to christen this ability to exploit older ideas in new arenas "Third Mover Advantage". Third Mover Advantage takes its name from the business school model of "First Mover Advantage", the supposed edge the market leader has in establishing a dominant position. Third Mover Advantage is the advantage enjoyed by companies who join the fray after the original players have created a market, spent all their marketing dollars grabbing market share, and lowering their prices to the minimum competitive level. At that point, any third company who can support the feature set of the first two can enter a mature market, spend a fraction of the marketing dollars and, best of all, not have to support the kind of brittleness and feature creep that often results from heated combat between companies.

Excel enjoyed third mover advantage after VisiCalc and Lotus1-2-3 created the market for spreadsheets, and third mover advantage may be at work in the browser market today, with Sun Microsystem's HotJava browser poised to enter the market created by the larger participants. If I am right in identifying this as a serious market force in the network computing world, then the eventual winners will have to include the companies who are not now involved, but who will jump in when the market has settled on some set of standards. Far from a winner-take-all outcome, this will be winner-take-some, where the companies who successfully set the standard will profit, but at the expense of increasing rather than decreasing the amount of competition they face from later entries in the market.

More generally, if two ideas or models come into conflict, at least where open systems reign, the struggle tends to realign more than destroy (mainframes are now used for what mainframes are best at, not for everything), to preserve the losing ideas virtues in other media (Tetris runs on Pilots, vt100 is now a software emulation mode of largely ignored hardware) and to engender the next round of competition for winning ideas. With equilibrium as a model instead of winner take all, the first question you should ask yourself when some idea seems ascendant is "What sort of competition will success open this idea up to?" and the first question you should ask yourself when some idea seems ready to be consigned to the dustbin is "When will this be back?"