Question

I’m slowly learning WPF using this article and other resources.

I am focusing on the application logic – defining the model + viewModel, and creating commands that operate on these. I have not yet looked at the view and the .xaml format.

While I am working on the logic, I want to have a view that can render any viewModel I bind to it. The view should

• Render any public string properties as text boxes, and bind the text box to the property
• Render the name of the property as a label.
• Render any public ‘Command’ property as a button, and bind the button to the command (perhaps only if the command takes no arguments?)

Is something like this possible while maintaing the MVVM design pattern? If so, how would I achieve it? Also, the article suggests to avoid using .xaml codebehind – can this view be implemented in pure xaml?

Answer 1

I don’t think it is possible in XAML only. If you want to generate your views in runtime then you have to just use reflection over your ViewModels and generate controls accordingly. If you want to generate views at compile time then you can generate xaml files from your ViewModels at build time with some template engine (like T4 or string template) or CodeDom. Or you can go further and have some metadata format (or even DSL) from which you will generate both models and views and so on. It is up to your app needs.

And also in MVVM code-behind is Ok for visual logic and binding to model/viewmodel that can’t be done in XAML only.