Should The Next Windows Be Built On Linux?

scrm writes "The next version of Windows should be built on top of Linux, according to this article by Robert Cringely of PBS." If Microsoft wanted to, they could be the world's largest vendor of Free software .. couldn't they?

Comments

Pardon?

[Acidic_Diarrhea]

"Now back to Microsoft putting Windows on top of Linux. Linux is better, faster, stronger than whatever is living underneath XP now, right?"

This article posed an interesting proposition but in order for the whole premise to fly, this snippet really has to be a "Well, duh" type of question. I really don't know that it is. Without any definitive proof of this, the author is merely picking at straws. How does anyone know where the clunkiness of Windows comes from? I mean, the NT kernel isn't exactly a slouch and I'm not sure that Linux is vastly superior. How are we to know that the windowing system isn't the problem? Perhaps the real Operating System-type services provided by the kernel are faster and more powerful than the ones provided by Linux. We just don't know because a separation of the windowing system from the real OS isn't possible with MS's closed source system. Thus, this argument isn't really credible to make. It's an interesting hypothesis but there's nothing solid to say, "Yes! You're right!" There's really no way to know which underlying OS services are better provided by XP or Linux.

Re:Everyone is missing the point.

[jpmorgan]

No, Cringley is missing the point, and he's talking about something he doesn't understand: namely the Windows/NT mix.

In some respects he's right, but accidentally (a stopped clock is right twice a day), in that Windows is built on top of another operating system, in this case, NT. But to transition to another base there are three questions that would have to be answered:

• Is it possible?
• Would it be better?
• Would it be economic?

Is it possible? Not without a lot of modification to Linux. NT is not UNIX and has a number of fundamentally different idioms; while Win32 abstracts a lot of this, it still pokes through in a few places. Even if Microsoft implemented features in the Linux kernel necessary, they'd still be forced to deprecate half the API and force developers to rewrite their applications to take full advantage of the new architecture. And if they rewrote Linux enough to make this unnecessary, it wouldn't be Linux anymore - it'd be an NT rewrite.

Would it be better? Cringley simply assumes that Linux is faster, more stable, etc... than NT. Windows is notorious for being unstable, although most of that reputation is due to the Win9x line. Win2K/XP have been known to crash on occasion, but unless you're using some seriously broken hardware, or have fucked its internals up a lot, it doesn't crash that often, and even then the vast majority of crashes are due to the Win32 layer, not NT itself. NT has a stronger security model, is realtime and fully reentrant. In short, the problems with Windows 2000/XP are not the fault of NT, but Win32 itself. Exactly how would porting Win32 to Linux solve these problem?

Would it be economic? The marginal benefit of porting to Linux would be minimal, and at great expense. I can't see how Microsoft would justify it.

Cringley suggested something that is fundamentally highly technical without understanding the real issues involved, which was stupid. This is particularly ironic when you consider the section of his site saying that people should listen to him since he knows what he's talking about. Once again this simply proves that he's nothing more than a digital snake-oil salesman - under the guise of holding an expert opinion, he tells people what they want to hear in exchange for ratings.