Surprisingly, I like the new programming model proposed by Microsoft for Longhorn.

The way I normally program interactive applications is by dividing my code into the trio: model, view, and controller. The view is responsible for displaying stuff; the controller, for processing user input; and the model, for the hard work.

In Longhorn (codename for next version of Windows), you'll be able to use three different languages for the three components.

The view in Longhorn is programmed a lot like a web page, using XAML (Extensible Application Markup Language) which is a lot like HTML. The simplest applications could be written entirely in XAML, just like most web pages are written in HTML. That's the view part.

If you want to add more interaction with the user, you have to add a controller. You do it, for instance, in C#. The interactive elements in your view will have C# handlers. This is again very similar to how interactive web pages are made--using some kind of scripting language, like JavaScript or VB.

Finally, if you want your application to do some heavy work, you'd write the model--and here you may choose yet another language, for instance C++ -- for robustness and performance.

I think it's a cool idea.