Windows Apps on Linux May Not Be Such a Great Idea
Written by Steven J. Vaughan-Nichols
Friday, 15 February 2002
As I write this, many otherwise sane Linux people I know are going gaga over the idea of running Win- dows applications on Linux with Lindows. "It'll make Linux the operating system for the desktop once they can get Office XP running on it," gushes one of my friends.
Oh, please! It won't do anything of the sort. First, while MP3.com millionaire Michael Robertson certainly has the bucks to pursue the dream of a cheap operating system that can run both Linux and Windows programs, it's vaporware now. I'm told that there are plans for a beta release by the end of 2001 and a shipping product in the first quarter of 2002.
Right, a company founded in September will have a shipping operating system by March 2002. I don't think so. Can they put a DOS/Windows emulator on Linux by then? Sure, the company is working with the Wine Windows 16/32-bit emulator, which dates back to 1993. However, even before Wine, DOS/Windows on Unix had already been done, over and over and over again.
We've had Microsoft operating system emulators and virtual machines for Unix since before Linux was around. I wrote my first review of DOS on Unix systems for Computer Shopper in 1990 using Interactive Systems VP/ix and Locus Merge.
Today I'm running NeTraverse's Win4Lin 3.0 (a Merge descendent) on Caldera OpenLinux Workstation 2.3, and I'm writing the first draft of this column in Microsoft Word for Windows 2.0c on it. If I really wanted to, I could get Office 97 running on it.
And you know what? It works well. Not only can I run Office, but I can also run some of my other favorite Windows applications, like Quicken and Photoshop. And if you don't care for Win4Lin or Wine, you can also just run multiple operating systems on one PC at once with VMware.
So you see, there's no need for anything like a Lindows. What it's trying to accomplish has already been done several times.
My point is that simply being able to run Windows applications on a Linux platform isn't going to make a significant impact on the Linux desktop world -- even if it's Office XP. It has never done so before, so why should it now?
Oh, I see how it can be useful. Sometimes you really do want a Windows application. Indeed, for some jobs like cross-platform development or technical support, being able to switch from Windows to Linux and back again on a single machine is a pretty useful option. But that's not where the market is. It never has been and it likely never will be.
The whole idea of emulation and VMs (Virtual Machines) has two fundamental flaws. The first is that try as they might, nobody can keep up with Microsoft's changes to their operating system and APIs. The slap on the wrist meted out by the Department of Justice and a slight opening of Microsoft's source code isn't going to change this. Neither a VM nor an emulator can run Windows and its applications as well as Microsoft's native system running on a standalone machine. The vast majority of people aren't going to come over to Linux when they know that only some Windows applications run on a Linux box. They'll stick with Windows.
The other flaw is that by wasting all this time (I'm talking years!) and energy playing catch up with Microsoft on its own playing field, there's less time and energy for native open source Linux applications. Every minute that a developer wastes trying to copycat Microsoft's operating system and APIs is a minute that could have been spent on some significant Linux application development.
As much as I like the emulators, they only hurt Linux's own software development. Even if we assume that due to a burst of programming genius Lindows can actually run all XP applications by next spring -- it won't happen, but let's assume -- then Linux application development will come to an end.
If you can run Office XP, then what need is there for StarOffice? We've seen it before. As Microsoft Office gained market share by hook or by crook in Windows, competitors like Corel WordPerfect and Lotus SmartSuite became mere shadows of their former selves. The same thing is likely to happen to Linux applications.
I don't know about you, but I don't want Linux to become a secondary platform for Windows applications. It may be Linux under the desktop, but it would only be Windows to users' eyes.
© 1999-2005 Infostrada Communications, LLC ()