And oh yeah, thanks for the great post too. I’ve been banging my head against a bunch of MVVM WPF examples and trying to understand what the hell the difference was between the implementations. Recognizing the difference between ICommand and RoutedCommand implementations was the key! Woo freakin hoo. Great post.

I suggest using CommandManager.RegisterClassCommandBinding().

Josh

Can you help me with Application wide commands?

Does any such term exists? What I want is, that I define CanExecute and Execute handlers on a application/top window/main frame level and all the child windows or pages of main frame pick those handlers automatically. Is this possible in any way?

If you got confused with what I want, I want something similar to what is being discussed here:
http://forums.msdn.microsoft.com/en/wpf/thread/cc2645a3-014b-4244-a91c-bbe40872cee0/

Regards,
Yogesh.

Josh

I was reading your article on WPF commanding and I have a question. I have an application that has a complex layout with multiple tab windows imbedded in a document container window which hosts a flowdocument. When I create a command for ‘cut’, ‘copy’ and ‘paste’ (from the ApplicationCommands class) and place it in a tool bar menu, the ‘tunnel and bubble’ fails to enable these commands when I select text in the flowdocument, but if I place the menu directly iside the xaml element that holds the flow document it works. Why does WPF fail to bubble up to the top-level toolbar but it works if I place the menu farther down the tree? Any idea on what to do to fix it?

Great article, and I agree the images are funny. I was interested in a scenario Kent mentioned “commands really shine when developing custom controls.” I was looking for a way to make commands for a component executed somewhere centralized away from the component, even away from the client implementing the component.

Scenario:
I have a Window [Win1] and it contains a component [Comp1]. Comp1 contains a menu item and it is bound to a central command. When executed Win1 determines whether it can be executed but the command executes in another class, like a service, and it works of some context. This would be ideal because I can then code this CommandController similar to a service for a specific context. This imaginery CommandController would have exposed commands and handlers for those commands. The final goal in this scenario would be that another component, possibly in another Window, could bind to the same command because it would want to deal with the same context identically.

Hope this make sense. I am still in search to make this happen and I believe your article helped me understand the usage of routed commands and my limitations.

Thanks again.