The most striking thing about Aaron Ricadela's excellent article about the future of Windows was the dramatic discrepancy between what Windows is and what it could (must) become.But to read about the utopian operating system Ricadela so deftly sketched out--secure, small, fast, and Webified--on the eve of when we're expected to implement one that possesses absolutely none of those qualities was extraordinarily frustrating. (I'm frustrated with Microsoft, not Ricadela, who does a terrific job of framing the relevant issues.)

All I could think was: Okay, Steve (Ballmer), Ray (Ozzie), and Steven (Sinofsky), I'm glad you're moving in the right direction, but what do we do now?

Every time we have to upgrade any kind of software, we have problems. Upgrade an operating system? Oy. Multiply that by thousands and thousands of desktops, and you have the dilemma of today's IT manager facing the Vista upgrade monster. He or she has to plan it, execute it, fix all the problems that come up, train people, and bolster the numbers (and morale) of the help desk to prepare for the many, many calls that will pour in. It's laborious and damned expensive, and in the end you have to ask, is it worth it?

Apparently, a lot of people don't think it is.

According to a poll published in early August by JupiterResearch, 50% of firms say they either have no plans to implement Vista or will wait at least 13 months after the operating system is released to do so. Astonishingly, 13% of the companies that responded to the survey had never even heard of Vista.

An ITWire poll from late July also asked enterprises about their Vista plans and came up with even more dramatic numbers: 64% said they had no plans to upgrade at this time.

But back to the article and to Microsoft's plans to create a "hybrid" operating system that would fuse the online and PC-based software worlds. It was interesting to read Rob Enderle's observation that "Netscape was absolutely right" about the rise of the Web, and that the desktop component of software could be easily "trivialized." It brought to mind Marc Andreessen's famous boast--was it really 11 years ago?--that the Web would reduce Windows to a "poorly debugged set of device drivers."

Clearly with Vista, Microsoft is doing its best to try and prove that prediction false. And Vista obviously delivers much, much more than Andreessen's scornful dismissal warrants.

But I don't trust Microsoft. I'm not talking here about its ethics, morality, or intentions, but about its ability to execute. That's the critical question. If we can't depend on Microsoft to deliver secure, reasonably bug-free software to our desktops, why should we trust that it can deliver secure, reasonably bug-free services over the Web? And presumably more will be at stake there if Microsoft makes good on Ozzie's promise to deliver processing power, data storage, and communications bandwidth to enterprises via the large data centers it's building. Imagine all our crown jewels in Microsoft's hands. Does this sound like a good idea?

Well, I've posed lots of questions here, but have precious few answers. Do you? If so, respond here.