Exactly ten years ago I was sitting in a small but cosy flat in the west of Helsinki, waiting to interview its owner. He was busy in the tiny kitchen, which lay just past the entrance hall decked out with dozens of cups and shields won at Karate competitions, preparing a cappuccino for each of us. As you've probably guessed, his name was Linus Torvalds - the trophies belong to his wife.

Wired's decision to devote the first mainstream feature to him and GNU/Linux was brave in 1996. At the time, free software was barely known outside a small, if growing circle of hackers. Linux was five years old, and still wet behind the ears. Although Apache had recently overtaken NCSA's software as the top Web server, it was one of the Internet's better-kept secrets. On the desktop, Windows 95 was unchallenged, as was Microsoft Office. Meanwhile, Internet Explorer was rapidly taking advantage of Netscape's continuing missteps in the browser sector to establish a near-monopoly there too.

Fast-forward ten years. GNU/Linux completely dominates the field of supercomputing. In the world of business computing, GNU/Linux is thriving in the server room, and is now the only serious competitor to Windows there. Apache's market share is hovering around the 60% mark; admittedly, this is its lowest level for four years, but that's not bad considering the fact that it is up against a company with a $30 billion cash hoard to buy friends and market share.

Even on the desktop, things are looking up for open source. Firefox has emerged from the ashes of Netscape Navigator to take ever-bigger bites of the browser market: according to one survey, it has over 20% in Europe and a respectable 13.5% in North America. And with OpenOffice.org and OpenDocument Format, the free software world finally has a standard that is able to take on - and beat - Microsoft Office.

These are extraordinary achievements; indeed, open source has become so successful, so pervasive, that we are in danger of forgetting how much has been done.

Against this background, then, I'd like to pose a question: what can't open source achieve in the next ten years? I phrase it this way, because it emphasises the fact that free software is likely to achieve much more than we might think. After all, who ten years ago would have been bold enough to predict that IBM - the archetypal conservative corporation - would place GNU/Linux at the heart of its strategy, or that the then-new Java would one day be released under the GNU GPL?

So, for example, I take it for granted that open source will be as successful on the desktop as it has on the server - with the caveat that the desktop itself may well be far less important in ten years' time. I also assume that everyone will be using ODF as the standard for document interchange and storage, and that GNU/Linux will consolidate its growing success in the field of embedded systems.

What's left? What won't be mostly running on free software in a decade? Or, to put it another way, what will Microsoft become in that time, as it finds its two principal revenue streams - Windows and Office - begin to dry up?

Microsoft's major product categories have followed a very similar trajectory. The first few iterations were dreadful, but eventually, through sheer dogged perseverance, the company has managed to produce winners that dominate their respective markets. This can be seen in the evolution of Windows on the desktop (3.0, 3.1, 95, XP), as well as server-side (Windows NT 3.1, 3.5, 4.0, 2000, Server 2003).

It can also be seen with the Xbox. Where earlier models were the clear underdog to Sony and Nintendo, there is increasing evidence that the Xbox 360 will emerge as the leading third-generation gaming platform. This is the area where open source is weakest: there are still relatively few free games, and no tools comparable to Microsoft's new XNA Game Studio. This means that Microsoft has more time to make money from its proprietary approach here than in any other field.

I suspect that we will also see the same pattern with Microsoft's Zune. By all accounts, the first iteration is awful; later models will doubtless improve until it becomes the dominant proprietary music player - not least because Microsoft can pay bigger incentives to the music industry than anyone, as it has already started to do. The Xbox 360's success and the company's ambitions in the field of home entertainment offer clear hints that this is where Microsoft's heartland will lie in the future, as traditional computer markets are increasingly commoditised by the inexorable rise of open source.

Glyn Moody writes about open source at opendotdotdot.