Unless you started with absolutely brilliant forward-thinking architecture and were uncompromising in adding features and code in each release, you eventually end up with a mess both inside and out. The pragmatic reality of commercial software (and perhaps even more so with open source), is that there are always compromises. Their effects compound over time.

At some point, you really need a fresh look at the whole space you’re in to understand the total user experience you want to create, to design a product, an interface, its underlying architecture and implementation. To start a new mess.

Yet the cost of doing so is enormous. While simple base functionality may be relatively inexpensive to create from scratch, the depth of support for all the little odd cases and special needs can be truly daunting. If your product is a platform, the complexity of ripple effects is mind-boggling. Plus, you’re abandoning a whole ecosystem when you make this kind of jump. In both Apple and Palm’s case, as they made the leap, the health of those ecosystems reduced the cost of leaving them behind.

With knowledge of issues brought up in and constrained by past versions (both in interface and internals), and freedom from starting over can come powerful simplicity, efficiency and effectiveness. Elegance in code and interface.

Which brings us back to the Windows platform and the Office platform. (Are they even two separate platforms?) Plus Windows Mobile.

(BTW, I loved that NeXT didn’t separate the word processor and spreadsheet from the OS. They were objects that any programmer could include anywhere in their interface. Regardless of how well that worked in that implementation, would another generation let us use data stored in the cloud and cached locally, wherever we log in, with whatever the local instances are for these interchangeable application objects, perhaps even swapping them out to suit our preferences, all abstracted seamlessly?)

What would be the point of a new Microsoft mobile OS?

Much of why Windows Mobile exists is to handle all the IT-department requirements around the world (as silly as any of them may be).

While it may be POSSIBLE to do so elegantly, is it PRACTICAL to do so? The level of complexity required to balance so many constraints may be beyond human capability to see the whole picture to create that elegant architecture, interface and implementation, or at least be prohibitively time-consuming and expensive.

MicroSoft in its current incarnation exists to make happy a whole ecosystem of corporate IT departments, hardware/software vendors, consultants and the rest. And to do so profitably.

To change the platforms fundamentally would require fundamental change to that ecosystem. There would be mutiny. That’s why Apple (and perhaps WebOS and others to come), are computers “for the rest of us.”