Reading Windows tea leaves

Though not exactly lighting up the blogosphere, speculations about future directions in the Windows platform have been rife. This article at APC compiled the various rumors together about Windows 7 (the successor to Vista...whatever name it ends up having) into a neat package, and this article on Ars Technica spoke of changes planned for Microsoft's Windows Mobile platform.

Keeping these blogs clearly in the "speculation" realm is the fact that Microsoft has been a lot more restrictive of information on future directions in its flagship platforms. Even though I work for Microsoft, those information controls are just as strong internally. So, unfortunately, I don't have access to "special information," and if I did, I wouldn't write about it (which is why I almost never write about IPTV).

That being said, some of the rumors are more concrete, and others have a ring of truth to them because they seem to deal with issues that most recognize as being critical to the future of the Windows platform.

The first issue, I think, relates to the "balkanization" of embedded development from desktop Windows development proper. That may be a somewhat strong word, as one of the competitive advantages of Windows CE is that it conforms for the large part to the wider Windows platform ecosystem.  The WIN32 APIs found in Windows CE and desktop Windows are mostly the same, and the version of .NET that exists for embedded uses mostly supports the same interfaces.

On the other hand, important differences exist, chief among them that the driver model is different between the two platform. Just unifying the model between embedded and desktop platforms would go a long way towards bringing the full range of hardware compatibility found on desktop Windows to the embedded space.

MinWin appears to be a solution to that problem.  Billed as a common bare-bones core that would be used across Microsoft's platform line (desktop and embedded), it could provide the kind of consistency that would exist if Microsoft ran the full Windows Vista in even small embedded devices...at least from a device driver standpoint.

The importance of this move has grown now that Microsoft's competitors are offering this kind of consistency.  Well, embedded Linux always had it, but as is now well-known, Apple uses stripped-down versions of Mac OS X in the iPhone and iPod touch...an incredibly smart move, in my opinion. Microsoft needs to do something similar while bringing its more robust software platform to bear on embedded devices (which it already, mostly, does). Such a move would open a device compatibility floodgate in ways that Linux and Apple's platform consistency, due to their smaller market shares, do not.

The second issue is one of UI, and though I think Microsoft does a better job with UI than some of its open source competitors (Microsoft fills a middle ground in that regard), I still think there is further to go.

Historically, Microsoft is a company that catered more to businesses than ordinary consumers. That worked wonders when computing was still mostly the domain of business. Today, however, computing has broken free of its business roots and now can be found in our music playback devices, our phones, our cars, and even in our entertainment products.

This is a difficult shift for a company that is, at heart, oriented around software platforms (though newer business lines, such as XBOX / Zune, are further along). Fortunately, there is plenty of room to make that shift, as the Microsoft platform with its supporting toolset is so far ahead of competitors (in my opinion, of course). At a minimum, good UI designers must be given the power to demand a certain look and feel in Windows.

That seems to be happening with desktop Windows. Steven Sinofsky, the new Windows VP who replaced Jim Allchin after he retired, has put Julie Larson-Green in charge of the Windows User Experience. Ms. Larson-Green was behind the UI changes for Office 2007, the product group from which Sinofsky sprang (which explains her new prominence in Windows).

I love the UI changes in Office 2007. Granted, there was some grumbling from the "change is bad" crowd, but anyone who spent more than an hour with the product quickly found that it made using Office 1000 times easier to do. I can find features the existence of which I had only heard rumors, which is what a good UI is supposed to do. It shouldn't be so complex that it hides the features that teams of developers worked hard to build into the software.

Ms. Larson-Green, to my mind, counts as a good UI designer, and given Sinofsky's track record, he is the kind of manager to give good UI designers the power to shape directions in major products.

Similar shifts seem to be occurring in Windows Mobile. I have gushed fairly often over the past year about the iPhone, because lets face it, the iPhone touch UI and simple task-based layout is an amazing leap forward from a device UI standpoint. The Windows Mobile standard UI, by contrast, is predictable and brand identifiable...and about as exciting as an accounting spreadsheet.

As part of Microsoft's move to make its products more consumer-oriented (a directive which comes straight from the CEO), updating the Windows Mobile UI was critical. As that Ars Technica article hints, upcoming releases are intended to simplify things considerably, and long-term plans show that designers have risen in prominence in a platform group that has, for the most part, concentrated exclusively on the enterprise.

All of this does not mean that Microsoft shouldn't continue to push hard on platform innovation (Volta is a good example). That momentum, however, would be hard to stop, as Microsoft has stuffed itself with computer science experts from around the world who aren't suddenly going to take up basket weaving because more emphasis is being given to good UI.

Learning the skills necessary for a more consumer-oriented product line is an enhancement to the way Microsoft has historically done things. That history is not one to be dismissed lightly, as it one that proved wildly successful. Continuing the core orientation laid down by Gates while adding new approaches designed to accomodate changed demand patterns seems a rational approach to change.

Platforms are still critical, but platform plus UI makes a market-beating combination.