Don,

While I think F# in Winforms 2.0 is useful, I'm really interested in F# and WPF. The reason that I mention the following points is that the effort to build up F# collateral for Winforms 2.0 usage in general form may be expensive and it would appear that WPF changes things dramatically.

Here are a few observations:

1. I haven't had a chance to dig in as far as I'd like, but one of the guys at work has done considerable research on WPF and is a maniac Winforms 2.0 guy. I've had a chance to look at some of what he's mentioned.
2. WPF seems to abandon much of the original premises of Winforms 2.0 for control containment.
3. Properties seem to be generated on the fly for controls in WPF.
4. Control containment determines the properties that are exposed in WPF. For example, if a control is placed in a grid, then the grid cell position is held for "location" properties. If a control is placed in a flow, then the absolute (relative to the container) x, y positions are used for "location" properties. The superset of properties for considered use for a control within a container is not built at design time.
5. Other characteristics of control USAGE will also determine what properties will be built for WPF usage at runtime.
6. WPF appears to make heavy use of a decorator/strategy pattern for these property emissions. Essentially, this means that you can write a control with regular event handlers that can take advantage of properties that don't exist at design time, but you know will exist at runtime.
7. There also appears to be heavier usage of Inversion of Control and Dependency Injection in WPF than there had been in Winforms 2.0.

I wish Sells & Griffiths would cover this stuff more in detail. I can only hope that Petzold does his usual magic and gives us a deeper picture of this stuff in his new book. But I will say that the Sells & Griffiths book is the ONLY book to date that has any WPF substance at all.

---O