I was exchanging e-mail recently with my friend Mike Class, SJ, who is associate dean of the Graduate School at Marquette University in Milwaukee, Wisconsin. Mike, who is a Jesuit priest and therefore naturally drawn to the whole idea of conversion, wants Microsoft to build its next version of Windows on top of Linux. And you know, it actually makes some sense!

The idea of Windows as an operating system is purely a product of the Microsoft marketing department, and not some law of nature. When Windows 1.0 appeared, it was a separate program you loaded on top of MS-DOS. Same for Windows 2.0 and 3.0 up through MS-DOS 6.22, the last standalone version of DOS sold by Microsoft. That way, they thought at the time, if Windows ever proved to be a commercial failure (this was far from certain and it is easy to claim Windows WAS a failure before 3.0), it would have been easy for Microsoft to punt back to MS-DOS.

It was only with Windows 95 that Microsoft decided that Windows was a success, dropped the second box and integrated the products. But are they really integrated? No. DOS 7.0 was under Windows 95 and DOS 7.1 brought the FAT32 file system to Win95, not the other way around. Even today, you can still get to a C: prompt under Windows XP, which means a disk operating system is hiding there no matter what Microsoft wants us to believe.

Windows XP is not an operating system. It is a windowing system that sits atop an operating system much as KDE or Gnome sit atop Linux.

The idea of running a windowing system atop an OS or even having competing operating systems under the same OS has been around for a long time. Back when Unix meant typing on a command line, in the PC world there were versions of DOS from vendors other than Microsoft, and in fact, some of those products are still available. IBM will still sell you a copy of PC-DOS 2000, and you can download a copy of DR-DOS 7.03 for free from DeviceLogics Inc., in Utah. Up through Windows 3.11 both these products worked as well or better under Windows than MS-DOS, and some people have claimed to have made them work under later Windows versions, too.

The history of DR-DOS is especially interesting because it went through so many hands. This MS-DOS-compatible operating system was written at Gary Kildall's Digital Research as a better version of DOS that would be Gary's revenge against Bill Gates, only it didn't work out that way, did it? Still, DR-DOS was a better product than MS-DOS at the time.

DR-DOS was eventually sold, along with the rest of Digital Research, to Novell, which already had its own DOS clone inside the early versions of Netware. Novell's version of DOS was noteworthy because of its Indexed TurboFAT file system, which was a response to the painfully slow performance of the Hierarchical File System built into MS-DOS 2.0. Indexed TurboFAT was a flat file system that was kept all the time in a RAM cache so that it was incredibly fast. Like all the really industrial-strength applications of the time, NetWare (like Autodesk and others) did whatever it could to get beneath MS-DOS. It was the only way to get high performance.

Novell eventually sold DR-DOS to Caldera, which at one point renamed it OpenDOS and started giving it away. Smart move, throwing away a brand name known by millions. What DR-DOS did for Caldera was give it an inherited anti-trust claim against Microsoft because Redmond kept changing Windows to make it incompatible with DR-DOS, which — if you do it just to be mean — is against anti-trust law. Caldera won more than $100 million from Microsoft in an out-of-court settlement, making their day, and then DR-DOS moved on to Lineo, a Novell/Caldera partnership for embedded software, and finally to DeviceLogics, which plans an 8.0 version for later this year, again aimed primarily at embedded apps.

Now back to Microsoft putting Windows on top of Linux. Linux is better, faster, stronger than whatever is living underneath XP now, right? Performance would improve. As Mike Class points out, by not having to develop its own OS, Microsoft could also save money. They wouldn't need however many people are presently devoted to maintaining the underlying OS that isn't supposed to be there.

And it wouldn't devalue Windows, precisely because Microsoft has done such a good job of making people think there isn't a DOS under there. Windows is the brand, and would remain so. And the nature of the General Public License is such that Microsoft would not be required to divulge much, if anything, about either Windows code or its applications, specifically because they would be sitting atop — not built into � Linux.

Apple has made a virtue of doing exactly this with MacOS-X, heralding its Mach kernel and BSD roots. Couldn't Microsoft do the same? The last I looked, Rick Rashid was still a Sr. Vice-President and head of Microsoft Research — the same Rick Rashid who, as a professor at Carnegie-Mellon Univerisity, was responsible for Mach in the first place. No biggie.

Then look at what this does to Microsoft's anti-trust situation. Suddenly, they aren't this overbearing monolith, but just another company pushing a windowing system and apps. True, they have a broader offering of window systems and more apps (not to mention more money) than any possible competitors — make that all possible competitors COMBINED. But that's not illegal. The Feds would simply go away, and even the current consent decree might no longer be required.

What this would do is level the playing field just a bit. The best windowing system and the best apps would win where they best meet the needs of users, which would vary from constituency to constituency. This wouldn't be the end of Gnome or KDE by any means, but what it would do is give 95 percent of the computers in the world a relatively standard and robust OS that could run a wide variety of windowing systems equally well. And for Microsoft, it offers that delectable prospect of selling Microsoft Office for Linux by inserting a bit of middleware to interface its apps with other windowing systems, again without having to adhere to the GPL.

It won't happen, of course, because Microsoft will want to maintain every advantage and would see this as giving-in. It would also have a negative impact on their language business, though that's not an absolute certainty. The part I love, though, is the idea of Bill Gates showing up at LinuxWorld to kiss Linus's ring.

And if it ever happens (the ring-kissing, I mean), don't forget Mike Class in Milwaukee, who came up with the idea in the first place. I don't know if Jesuit friars are allowed to accept big checks from corporations, but I'm sure Marquette University could always use the money.