Breaking up Microsoft isn't so hard to do

There is a revealing little comment tucked away in Microsoft's latest annual report. After all the wonders of its higher profits (up to $3.45bn, on revenues of $11.36bn for the financial year), it looks to the future. There, it notes that "as the desktop application market has become saturated, the sales mix has shifted ... to upgrade products. This trend is likely to continue."

In other words: our software isn't wearing out quickly enough. We can't get enough people to buy it. And there's worse: "Product upgrades ... have lower prices and margins than new products," the report adds. Now, is this bad news for Microsoft, or bad news for the rest of us?

Initially, it looks bad only for Bill Gates and friends. Unless, that is, the Microsoft can find a way to make us buy more expensive upgrades. Oh, is that Windows 98 coming around the corner? Do you find that your old version of Microsoft Word can't read files written by the latest version? Is there a millennium bug problem with your Microsoft application? Bill Gates has the answer: upgrade, upgrade, upgrade.

However, faced with that "saturated" application, there is another way that his company could regain its hold: wipe out the opposition. You might think that sounds extreme. Yet crushing opponents and swallowing up their market share is exactly how Microsoft has got where it is today.

A few examples. In the 1980s, Digital Research used to make DR-DOS, a rather good competitor to MS-DOS. Microsoft strangled it using traps in its early Windows product which generated worrying messages if you ran it on top of DR-DOS. Digital Research went down the tubes. Garry Kildall, its founder, committed suicide.

Around the same time, WordPerfect was dominant in the word-processing market. Gates calculated how much market share he had to skim off to disable that rival's cash flow, which in turn would make it impossible for WordPerfect to keep up the development of its product. The magic figure was 5 per cent. By dint of wild promotions (at huge cost) and hefty marketing, Microsoft pushed Word to 5 per cent market share. WordPerfect crumbled.

Which package is now dominant in word-processing? It's not WordPerfect, which, meanwhile, is on its third owner. Was it an inferior product? Not necessarily; it just didn't have the marketing resources (including the chance to be bundled with other Microsoft applications and operating systems) of Microsoft.

More recently, the tale of Netscape's Navigator and Microsoft's Internet Explorer has been told many times. By offering its Web browser for free, and forcing PC manufacturers to make a choice - either get rid of Navigator, or do without Windows 95 - Microsoft undercut a rival that was well ahead of it in development and quality.

Gary Reback, a lawyer waging a war against Microsoft, put the "icon battle" of the browsers thus in a recent interview with the Silicon Valley gossip magazine Upside: "If you view the desktop as a billboard, [Microsoft] owns all the billboards in the world."

But that's enough examples. The question is, what - if anything - should be done about it?

Those who argue that Microsoft should be left alone fall into two camps. The first tend to be computer manufacturers whose interests are best served by not upsetting Microsoft. The second are those who insist that "America is a free country, you're all socialists". (They also say that Microsoft is an "innovative" company. Readers are invited to name one product from Microsoft which does not copy an idea already generated by someone else. Note that even MS-DOS was itself licensed from another company.) The "land of the free" argument is weak, and getting weaker. Windows is now so dominant on PCs that there's an argument for splitting it off entirely from the corporation that has developed it.

There is a precedent. In the days when America's greatest products were made of steel, the US Supreme Court established the principle of an "essential facility". The case arose when a number of railroad companies banded together and bought a crucial railway bridge over a river. They then refused to let rivals run trains across the bridge.

Trains being generally bad at swimming, the stranded companies sued for relief. The court ruled that the bridge's owners had to grant everyone access to this "essential facility".

In the same way, if you can't get access to Windows, you'll have a hard time introducing your software to 97 per cent of the general public. On that basis alone, the US Department of Justice could argue that Windows is now an "essential facility".

The fact that Microsoft built this virtual bridge is neither here nor there. Someone had to build the bridge in the original example, too. But if you're going to declare Windows an "essential facility", open to all (perhaps by publishing the source code), might it not be simpler - and, in operational terms, more efficient - to go the whole hog and split Microsoft up?

It would be relatively simple to do. Let's define three companies: Microsoft Operating Systems, Microsoft Applications, and Microsoft Media. Each would be publicly quoted and owned; existing US laws would prevent wide cross- ownership. Separate staff, separate locations, separate ownership.

But, you say, what is an operating system? What is an application? Second question first. An application runs on an operating system. Anything that Microsoft has offered on more than one computing platform such as DOS, Windows 3.1, Windows 95, WindowsNT, IBM's OS/2 or Apple's Macintosh operating system must be an application.

That would put Internet Explorer firmly into the "application" bracket because Microsoft makes a version for the Mac. Ditto for Word and Excel, its word-processing and spreadsheet products. So too PowerPoint, its presentation package. This sort of winnowing process would rapidly designate the applications. "Suites" of packages also help to define applications.

Where does this leave the operating system, then? After all, Stac Applications developed a "file cruncher" that would save space on your hard disk. Microsoft appropriated it (illegally, as a court case later decided) for MS-DOS 6.0. Is Stac's technology an application, or part of an operating system?

Well, clearly it's an application that has become subsumed into an operating system. Yet it could still function as a stand-alone application; you might even be able to find it running on 286-based computers (they do still exist, you know; I have one which used to "run" Windows 2.0, in the same way that treacle "runs" uphill).

In the world of a breakup, the two Microsoft companies, Apps and OS, would have to bid against each other if they wanted to resell Stac's technology. Who gets the reward? Stac, for innovating. But doesn't it put up the price of your operating system, or applications? Yes, it does. But that's how free markets are supposed to work - consumers pay for extra utility while innovators are rewarded.

In general, the OS company would find it advantageous to tell applications programmers how to write code for its systems. Microsoft does this already - but its own applications programmers now have the advantage of direct access to insider knowledge. That wouldn't be allowed after a breakup.

Let's imagine what would have happened if this regime had been in place a couple of years ago, and Microsoft OS had decided that it would be a smart move to offer a Web browser in later versions of Windows 95.

What would it do? It could write one from scratch, but that would take time and cost a lot of money - which it wouldn't have so much of, being cut off from the revenues streams of applications like Word and Office97. It could get a company like Microsoft Apps to write one - but since it hadn't started, that would also take time and cost more money. Or it could turn to Netscape, which would also have versions of its browser for other platforms. Microsoft OS could license Netscape's browser and offer it in its operating system.

How odd. Netscape is reaping the rewards of being innovative, as you would expect in an efficient market, rather than being crushed underfoot in this one. Note that in the "breakup" scenario, you still get a world where one browser predominates. You do get a "standard", at least on the PC. But Netscape can't rest on its laurels (it would be foolish to), because rivals will be trying to beat it, and there's nothing to stop Microsoft OS from dropping Netscape in the next release of its operating system in favour of, say, Internet Explorer from Microsoft Apps.

And Microsoft Media? That's easy to define, too - applications of applications. Hotmail, which Microsoft recently bought, is a media product - a free e-mail service. If you can access it without even using a Microsoft company product, it must be media. (The Internet generally counts as media in that sense.) Microsoft has made a number of moves into the media world of late, including its "Web news" alliance MSNBC.

One lingering question is how big these three companies would be. Microsoft refuses to discuss the income or revenues from its different divisions. Any figures here would be only a guess (though the Media company would certainly be tiny). But those words quoted earlier from its annual report indicate that the applications market is not growing quickly. MS Apps might be the lame duck of the trio in cash-flow terms, though not in raw revenues.

Is this all just fantasy? Not quite. There are signs that Microsoft is getting ready to perform this breakup operation on itself, without the help of the US Department of Justice. Bill Gates has appointed three executives to the Microsoft "Office of the President". Hmm. Three people to fight over the post if Bill gets run over by a bus? Or three people to head up a split leviathan, which would co-operate where necessary?

One thing is sure: if we can think of it, then Bill Gates certainly can. The only questions are, what has he decided to do? And when will he do it?