Microsoft is a company with a ton of talent and resources, but essentially no unified direction. There are too many fiefdoms and silos. Ray Ozzie is bringing a new vision, but he is too quiet and too many folks simply don’t understand where his efforts are leading. To make matters worse, nothing is changing on the legacy side.

Take Vista as an example. Vista is the right OS at the wrong time, in many ways. Beleaguered by internal politics, maligned by the media, leveraged gleefully by MSFT bashers and rivals (Apple couldn’t buy this kind of press) over hyped and underdelivered, it is a shining example of what has gone wrong in Redmond. As the world shifts towards Netbooks and web services, Vista is an OS designed for that brief era of the escalating hardware arms race. MSFT should know by now that tech industry time runs like quicksilver and that a 2002 paradigm is long dead by 2006. Windows Mobile is a far bigger disaster and frankly too much of a topic to discuss. Only on the server side are things being done right.

So how can this be fixed? Personally I say it is time to scrap Windows, Windows Mobile, Win CE and Windows Embedded. A new Windows should emerge built on a small tight kernel that is portable. Dave Cutler is still there and Mark Russinovich has joined him. These guys can do anything. The minwin work is a step in the right direction. A tight Windows kernel that can move from x86 to ARM to MIPS to PPC easily would give MSFT massive agility. APIs and services would then layer on top based on the device capability and use case. The user experience component could easily be delivered in a consistent manner as long as Silverlight and a highly optimized .NET CLR exist for the device. Pay someone external if need be to create one photo experience, one media player and one browser that run quickly and look the same whether you are on a phone or on a $7000 desktop. These are commodity services and should be as dependable and predictable as AC power or legacy land lines for people.

The last piece of the puzzle is to bring in Rays vision. Cloud enable all of these platforms. Let the OS determine its level of connectivity, hardware capability and service entitlement and then transparently present the user with integrated cloud options that mesh seamlessly into the UI. To enable others to play, base this cloud plugin standard for the OS on well documented interfaces and make plugging in very open. So yes, the users bubbly cloud icons might, behind the scenes, be pointing them to Amazon or Google or even Apple.

Unfortunately I suspect none of this is likely to happen. What seems more likely is that Win 7 will be a well optimized revisit of Vista that stops the bleeding and Win Mo 7 will very possibly be the mobile version of Vista. Too little, too late and an anachronism in the face of competition that made it to market first and changed the paradigm before it ever saw the light of day.