Microsoft Windows turned 20 this week. Like any 20 year-old, Windows is heading into its third decade with a swagger in its walk and a hint of arrogance in its eyes. And with good reason: The road Windows has traveled for so many years looks just as familiar as ever and still promises to take it exactly where it wants to go.Like any 50 year-old, Bill Gates knows better than to believe for a moment that Windows will escape its share of twists, turns, tight squeezes and near-misses -- or that, in fact, Windows will even survive the journey. That's why I would like very much to know what Gates was thinking, if not saying, when MIT Media Lab chair Nicholas Negroponte showed him a working prototype last week of a sub-$100 laptop.

The $100 laptop is the product of the Media Lab's non-profit One Laptop Per Child (OLPC) project, which Negroponte had just announced in January, at the 2005 World Economic Forum. This is a project which, if all goes well, intends to place at least 150 million of its rugged, hand crank-powered laptops in the hands of school-age kids in deveoloping nations over the next several years. It's an initiative that, with its global reach, educational focus, and emphasis on self-empowerment, resembles Bill and Melinda Gates' own charitable ventures. Yet it's also a project that relies upon Linux and other open-source software -- for reasons that I suspect Gates understands, accepts, and tacitly, if not overtly, supports.

The economics of the project are important, of course, but only to a point: For Negroponte and his colleagues, open-source software is far more than a cost-management tool, even if their $100 laptop still (but likely not for long) costs $118 to produce. The project's backers know all too well that if they manage only to create millions of new software users -- in other words, consumers -- then they miss an important opportunity

If, however, One Laptop Per Child turns enough of these kids into producers -- software engineers, entrepreneurs, innovators -- whose "means of production" lie almost entirely between their ears, then the project will succeed. In fact, it could be the greatest success of its type, in terms of return per dollar invested, since the Marshall Plan rebuilt Western Europe more than 50 years ago.

If the project succeeds, it will also have beaten the odds: The typical "foreign aid" effort -- such as the IMF-backed heavy industrial projects of the 1960s or the Monsanto-fueled disasters packaged and sold as the "Green Revolution" -- create nothing, in the long run, except distorted markets and crippling public debt. Whether the donors' true motives were well-meaning or cynically self-serving, foreign aid to developing nations often fails even to leave recipients better able to fend for themselves, much less to envision a future where they can aspire to something more than survival.

Yet this is precisely what One Laptop Per Child hopes to achieve. Each of these laptops delivers the raw material -- the compiled software, the source code, and an opportunity to use both in new and interesting ways -- that can enable these kids to learn, imagine, create . . . and, in time, to contribute and prosper.

You can see why, if Bill Gates offered tomorrow to put Windows on every one of these laptops, at no cost and in perpetuity, the project's management would have to say, "no thanks." That's precisely what they did, by the way, when Steve Jobs made the same offer with OS X -- an operating system that combines an open-source core (the superb FreeBSD) with the highly proprietary "Aqua" user interface. These two products, encumbered with restrictive licensing and business models, simply can't offer the same opportunities to the students who will use these computers.

In a broader sense, this dilemma brings Microsoft face to face with one of the most difficult and dangerous business challenges it will face in the years to come. Whatever the specifics of a given situation, it will always boil down to a single conundrum: In developing nations, Windows cannot and will not compete against Linux strictly on business or technological terms. Instead, Microsoft must guide Windows, and its other products, through a maze of shifting political and cultural hazards which very few westerners even fully recognize at this point, much less understand.

In some cases, a developing nation's government will settle this debate: China's public sector, including its schools, are already adopting Linux at the behest, overt and otherwise, or the nation's central government. In others, ironically, Microsoft's own fight against software piracy may settle the matter: Most Chinese firms that kick the pirated-software habit, for example, seem far more inclined to learn Linux, rather than pay for Windows. In still other cases, such as Brazil and Peru, government policy and public sentiment have aligned to support the use of Linux and open-source software as a sound long-term goal, often in the face of Microsoft's persistent short-term incentives (and threats) to convince them otherwise.

On the whole, it's easy to conclude that Microsoft lacks any coherent plan to deal with these typs challenges. In fact, if the company's disastrous Chinese operations, its business-as-usual platitudes that it uses to counter the Latin American open-source challenge, or even Steve Ballmer's Ugly American-esque software patent threats last year are any indicator, the company still doesn't have a clue about what's happening or what's at stake. It's a glaring, entirely fixable, and potentially lethal gap in the company's strategic vision.

So, will Gates support One Laptop Per Child -- and by extension, its embrace of open-source software as a tool for economic self-empowerment? As I mentioned before, I'm guessing that he will, and in a substantial way.

Yet there's a twist to my prediction: There is no guarantee that Gates and Microsoft will be of a single mind on this matter, even if both parties take great pains to play down this fact. The fact is, Gates may define his legacy more in terms of his philanthropy than through his role creating Windows and Microsoft, and sooner or later, both his words and deeds may reflect this fact..

Windows has, in any event, enjoyed a healthy and successful first 20 years -- its "father" saw to that outcome during almost every waking moment of his life. If Windows expects to survive another 20 years, much less to thrive and to stay relevant, its guardians --including those to whom Gates entrusts his business and technological legacies -- must learn to find their way in a business world where that old, familiar-looking road is likely to lead them straight off a cliff.