Gartner says Windows is "collapsing." Well, sure. Strictly speaking, Windows itself isn't in a state of collapse -- Windows XP is still useful to a huge population of customers. But for Vista and the Windows franchise as a whole, things do not look good.
By coincidence, 15 years ago this month, Lou Gerstner arrived at IBM.
When he got there, Gerstner found a company that literally didn't believe in its own future. The mainframe business -- the core of IBM -- was collapsing. Other business units were busy trying to turn themselves into stand-alone companies that could be spun off. The big blue ship was sinking, and everyone wanted off.
Gerstner turned it around. The specific tactics are interesting: a shift to services as the primary business, a deep dive into the Internet and Linux, and a vow that IBM would no longer try to lock customers in. Mostly, though, those tactics were incidental.
What mattered was this: Gerstner led IBM to change. He had to. He understood that IBM's old way of doing business just wouldn't work any longer. With a plummeting stock price and 100,000 laid off, change was the only option.
And it worked.
Now consider Microsoft. For years, the cliche has been that Redmond is the new Armonk. And just as IBM faced disintegration when the mainframe business collapsed, Microsoft faces disintegration if Windows collapses.
But the parallel runs deeper. Far too much of Microsoft simply doesn't believe in its future.
The future is change. That's what makes it different from the now. But Windows is in trouble -- and Microsoft is at risk -- specifically because Microsoft won't change.
Microsoft's product design philosophy is still "ever bigger, ever more bloated." That worked when Moore's Law made CPUs and memory ever faster. But now the market opportunities are in phones, games and low-end laptops. Bigger isn't better -- and Microsoft isn't changing.
Microsoft's software development approach -- nightly builds and daily bug fixes for years on end -- has been broken for a long time. It just won't scale for a code base as huge as that of Windows. There are other ways to do it -- but Microsoft isn't changing.
Here's the worst part: Microsoft's approach to customers has always been old-school IBM: Lock 'em in. But the first rule of lock-in is this: Never break your customers' software. And the second rule is this: Never force a product on customers that breaks their software when the previous version still works fine.
Vista breaks applications. It breaks device drivers. It breaks the strongest reasons for customers to stay with Microsoft. It actually threatens to break Microsoft's customer lock-in.
That's the one change Microsoft can least afford. And it's the change Microsoft is actually plunging headlong into.
That may sound appealing to us in corporate IT -- after all, we hate to be locked in. But we really do need vendors who want to lock us in. When vendors stop caring, we stop getting anything useful from them. We certainly need that from Microsoft.
And we need Microsoft to find new ways of designing and developing Windows -- radical reinventions of Windows, perhaps, but reinventions that don't break our apps, our infrastructure or our users' training.
Most of all, we need someone to lead Microsoft into all this change. Someone who understands that Microsoft's old ways are collapsing. Someone who understands that the future is the only option.
Let's hope Microsoft starts looking for that leader soon. Lou Gerstner isn't available.
Frank Hayes is Computerworld's senior news columnist. Contact him at frank_hayes@ computerworld.com.