For many applications, the ability to extend the application with third party features is essential. Microsoft's CLR Add-In team has been working on a formal model and API to make this task easier. Features include a standard way to handle discovery, loading and unloading, isolation, and communication with the host application.
On of the key features in the Visual Studio 2008 Beta 2 version of the Add-In model is the ability to isolate GUI elements. Add-Ins can now create their own GUI elements in a separate AppDomain that are displayed amongst the main content. So while the application sees the add-in as separate, the user gets a seamless experience.
The ability to use AppDomains is essential when working with code that is not fully trusted. By segregating the code, it can be run with a restricted set of permissions than the rest of the application. Not only does this help when dealing with potentially malicious code, it is helpful when the add-in code has reliability issues that would otherwise affect the rest of the application.
Jesse Kaplan goes into the technical details of how AppDomain Isolated WPF Add-Ins work in the CLR Add-In Team Blog.
