Just reading the roadmap from DevExpress
http://www.devexpress.com/Home/Announces/Roadmap2008.xml it is pretty clear that that sales for Windows Forms are still very high. This has meant that investment in the Windows Forms platform is to continue.
The principal problem with WPF for LOB applications, is that you need to pay for a designer, and they do not come cheap. In a six month turnaround application, that can easily be £60,000, or much much more. With most LOB applications, you purchase presentation layer components from Infragistics, DevExpress, Component One etc, and your developers thrash out the necessary business logic with the Office 2007 UI or whatever UI is flavour-of-the-month.
Because Microsoft have just created the fluent interface in the Office 2007 ribbon, I cannot see a significant change to this UI in the next office version. Office is still written in C/C++ and I don’t see a re-write for the next office version. As a result most applications written in windows forms today will continue to be relevant. Look at Visual Studio, the toolbar components are pretty much the same as Office 2003, even down to the icons that have not been changed for years. Visual Studio is a tool, and most LOB applications are tools.
Differentiated UI is a big risk in LOB applications, because you usually have a workforce with varying IT skills. You can safely assume that most users are familiar with Outlook, so creating an application with a similar look and feel lessens the learning curve. WPF opens up new doors, and for me in LOB applications, only the subtle use of it with nice 3D charting or reports is the way to go. If you move to far away from the familiar, you will end up with users struggling to learn the application. The net resultant is increased training costs for the application, for no real business value.