WPF control-derived class has a composition of non-WPF control-derived class. The latter class, say Inner, should trigger an Outer class WPF DependencyProperty change.
Essential invariant is that my Inner class shouldn't force any WPF namespaces to use whereby, Inner is part of API of a dll I am creating. So Inheritance from any WPF class is not appropriate solution.
What could be the best (or at least arbitraty) way to implement a callback from non-.NET derived class to a WPF-Control class to change a DependencyProperty object.
Some hypothetical callback mechanisms found:
c# delegates and c# events.
Can't solve this, becouce it is thrown a runtime exception
The calling thread cannot access this object because a different thread owns it.
when a callback mechanism tries to modify Background property (a DependancyProperty object I think is the key, or at least .NET defined property).
I guess that c# delegates doesn't work in single-threading manner - however, i didn't find this fact while reading c# documentation.
(One solution is through .NET Dispatcher class and Invoke(), but it seems too arbitrary and possibly slow - isn't any more WPF-friendly way?)
I had read about Routed events and DependencyProperties
in WPF documentation (however had't tested this yet), but it seems Routed event cant help (because my Inner class isn't part of a VisualTree) and I should't inherit from DependencyObject.
I read (I suppose) everything relevant to my problem. I'm sorry if some similar kind of question had been asked, I would be glad if you direct me to that thread, if the problem essence is the same.
Simplified example:
using System;
using System.Collections.Generic;
using System.Linq;
using System.Text;
using System.Windows.Controls;
using System.Windows.Media;
using System.Timers;
namespace WpfApplication4
{
internal class Rect:ContentControl
{
Inner inner = new Inner();
public Rect()
{
inner.Rotate += new Inner.RotateEventHandler(rotated);
}
public void rotated(object sender, RotateEventArgs args)
{
Background = Brushes.Blue; //Run time error
VisualTransform = new RotateTransform(args.Angle); //Run time error
}
}
public class Inner
{
Timer timer = new Timer(200);
public Inner()
{
timer.Enabled = true;
timer.Elapsed +=new ElapsedEventHandler(elapsed);
}
public delegate void RotateEventHandler(object sender, RotateEventArgs args);
public event RotateEventHandler Rotate;
public void elapsed(object sender, ElapsedEventArgs e)
{
RotateEventArgs args = new RotateEventArgs();
args.Angle = args.Angle + 45;
OnRotate(args);
}
protected void OnRotate(RotateEventArgs e)
{
if (e != null)
{
Rotate(this, e);
}
}
}
internal class RotateEventArgs
{
private double angle = 0;
public double Angle
{
set { angle = value; }
get { return angle; }
}
}
}
You can make use of DispatchTimer instead of Timer. Have the DispatchTimer in the Rect class itself.
Related
I have an application that allows users to move forward and backward
And the radio button is implemented like following:
For the first 4 times, I change the selected item for the radio button and then move next and then come back and do the same, the radio buttons work fine and the model is updated correctly as its GUI. However, after the first 4 times moving forwards and backwards, the radio buttons do not update its GUI.
Following is the navigation system:
Model and view are bound based on Next/Back button
Check boxs works fine!
Please give me some suggestions.
Thanks in advance
I finally found the answer:
If you have tried to bind the RadioButton鈥檚 IsChecked property in WPF to an object, you have most likely experienced the following problem: In OneWay bindings it works great. But if you have more than one RadioButtons binded TwoWay and you click on an unchecked one, you were expecting that the object to which the previously checked RadioButton was binded to receive the value of False. But you were wrong in your expectations. That鈥檚 because for some reasons Microsoft does not obey bindings and does not pass the False value to the DependencyProperty and instead of that they just assign the value False directly to the property, which ruins the binding.
There are many proposed solutions to this around the internet, problem with all those is that they do not work with dynamically generated controls. So since I had to find a way to make this working with dynamic controls, decided to make a wrapper of the real RadioButton which will correctly Bind in two ways. Here is the code for the wrapper:
using System;
using System.IO;
using System.Printing;
using System.Text;
using System.Windows;
using System.Windows.Controls;
using System.Windows.Controls.Primitives;
using System.Windows.Documents;
using System.Windows.Input;
using System.Windows.Interop;
using System.Windows.Media;
using Microsoft.Win32;
namespace Controls
{
public class RadioButtonExtended : RadioButton
{
static bool m_bIsChanging = false;
public RadioButtonExtended()
{
this.Checked += new RoutedEventHandler(RadioButtonExtended_Checked);
this.Unchecked += new RoutedEventHandler(RadioButtonExtended_Unchecked);
}
void RadioButtonExtended_Unchecked(object sender, RoutedEventArgs e)
{
if (!m_bIsChanging)
this.IsCheckedReal = false;
}
void RadioButtonExtended_Checked(object sender, RoutedEventArgs e)
{
if (!m_bIsChanging)
this.IsCheckedReal = true;
}
public bool? IsCheckedReal
{
get { return (bool?)GetValue(IsCheckedRealProperty); }
set
{
SetValue(IsCheckedRealProperty, value);
}
}
// Using a DependencyProperty as the backing store for IsCheckedReal. This enables animation, styling, binding, etc...
public static readonly DependencyProperty IsCheckedRealProperty =
DependencyProperty.Register("IsCheckedReal", typeof(bool?), typeof(RadioButtonExtended),
new FrameworkPropertyMetadata(false, FrameworkPropertyMetadataOptions.Journal |
FrameworkPropertyMetadataOptions.BindsTwoWayByDefault,
IsCheckedRealChanged));
public static void IsCheckedRealChanged(DependencyObject d, DependencyPropertyChangedEventArgs e)
{
m_bIsChanging = true;
((RadioButtonExtended)d).IsChecked = (bool)e.NewValue;
m_bIsChanging = false;
}
}
}
So now all you have to do is to use the ExtendedRadioButton instead of the built-in one and bind to the IsCheckedReal property instead of the IsChecked one.
Enjoy 馃檪
Has anyone ever heard of implementing IDependencyObject instead of inheriting from it -- that way one could actually create a class hierarchy instead of having to use only interfaces when trying to get both dependency object/property and custom behavior on our classes.
I want to have a hierarchy of class kinds that are directly usable in the context of an existing structure, i.e. Polygon. I want to be able to use my PolyType in any place, and without any more dialogue and indirection that would be required if I place the PolyGon existing type as a Part of my DependencyObject. But I also want to be able to have my class as the a) the target of {Binding} markup extension, b) Animate properties of PolyType and c) apply themed styling to PolyType.
I want to implement IDependencyObject instead of being forced to inherit from it directly, and obstructing my ability to be a direct descendent and usable in place of, PolyGon.
Not sure why you have to inherit from DependencyObject. I use a custom code snippet that generates the following code to register a dependancy property:
public partial class UserControl1 : UserControl
{
public static DependencyProperty MyPropertyProperty = DependencyProperty.Register("MyProperty", typeof(Polygon), typeof(UserControl1), new FrameworkPropertyMetadata(new PropertyChangedCallback(MyProperty_Changed)));
public Polygon MyProperty
{
get { return (Polygon)GetValue(MyPropertyProperty); }
set { SetValue(MyPropertyProperty, value); }
}
private static void MyProperty_Changed(DependencyObject o, DependencyPropertyChangedEventArgs args)
{
UserControl1 thisClass = (UserControl1)o;
thisClass.SetMyProperty();
}
private void SetMyProperty()
{
//Put Instance MyProperty Property Changed code here
}
public UserControl1()
{
InitializeComponent();
}
}
As you can see the DependencyObject can be any type of object. If this is not what you need, please post you code examples, or explain your situation better.
I'm developing MVVM application and I'd like to create global shortcuts for my application. It's necessary to navigatate application between modules. I have main window with code:
<UserControl.InputBindings>
<KeyBinding Command="{Binding ChangeModuleComand}"
Key="M"
Modifiers="Control"/>
</UserControl.InputBindings>
It works when I have focus on my window,
But there is a problem when i have focus on TextBox in that UserControl (Command is not fired).
Is that possible to catch that key pressure without lots of modifications in whole application?
A.
It sounds like your event is being handled before it bubbles up that far. Next I am wondering if you could move your input binding up to your window rather than a user control. My final option would be to remove your input bindings and create a global check in code.
Since you're already working with input bindings, I have added the code option below:
//In App.xaml.cs
protected override void OnStartup(StartupEventArgs e)
{
base.OnStartup(e);
EventManager.RegisterClassHandler(typeof(UIElement), UIElement.KeyDownEvent, new RoutedEventHandler(GlobalClick));
}
private void GlobalClick(object sender, RoutedEventArgs e)
{
var args = (KeyEventArgs) e;
if (args.KeyboardDevice.IsKeyDown(Key.M) &&
args.KeyboardDevice.IsKeyDown(Key.LeftCtrl) || args.KeyboardDevice.IsKeyDown(Key.RightCtrl))
MessageBox.Show("Works");
}
Documenting this answer for others, as there is a much simpler way to do this that is rarely referenced, and doesn't require touching the XAML at all.
To link a keyboard shortcut at the Window level, in the Window constructor simply add a new KeyBinding to the InputBindings collection. As the command, pass in your arbitrary command class that implements ICommand. For the execute method, simply implement whatever logic you need. In my example below, my WindowCommand class takes a delegate that it will execute whenever invoked. When I construct the new WindowCommand to pass in with my binding, I simply indicate in my initializer, the method that I want the WindowCommand to execute.
You can use this pattern to come up with your own quick keyboard shortcuts.
public YourWindow() //inside any WPF Window constructor
{
...
//add this one statement to bind a new keyboard command shortcut
InputBindings.Add(new KeyBinding( //add a new key-binding, and pass in your command object instance which contains the Execute method which WPF will execute
new WindowCommand(this)
{
ExecuteDelegate = TogglePause //REPLACE TogglePause with your method delegate
}, new KeyGesture(Key.P, ModifierKeys.Control)));
...
}
Create a simple WindowCommand class which takes an execution delegate to fire off any method set on it.
public class WindowCommand : ICommand
{
private MainWindow _window;
//Set this delegate when you initialize a new object. This is the method the command will execute. You can also change this delegate type if you need to.
public Action ExecuteDelegate { get; set; }
//You don't have to add a parameter that takes a constructor. I've just added one in case I need access to the window directly.
public WindowCommand(MainWindow window)
{
_window = window;
}
//always called before executing the command, mine just always returns true
public bool CanExecute(object parameter)
{
return true; //mine always returns true, yours can use a new CanExecute delegate, or add custom logic to this method instead.
}
public event EventHandler CanExecuteChanged; //i'm not using this, but it's required by the interface
//the important method that executes the actual command logic
public void Execute(object parameter)
{
if (ExecuteDelegate != null) //let's make sure the delegate was set
{
ExecuteDelegate();
}
else
{
throw new InvalidOperationException("ExecuteDelegate has not been set. There is no method to execute for this command.");
}
}
}
I'm sure this can be used for other controls as well, but haven't tried it.
I'm trying to implement drag and drop functionality in a Surface Application that is built using the MVVM pattern. I'm struggling to come up with a means to implement this while adhering to the MVVM pattern. Though I'm trying to do this within a Surface Application I think the solution is general enough to apply to WPF as well.
I'm trying to produce the following functionality:
User contacts a FrameworkElement within a ScatterViewItem to begin a drag operation (a specific part of the ScatterViewItem initiates the drag/drop functionality)
When the drag operation begins a copy of that ScatterViewItem is created and imposed upon the original ScatterViewItem, the copy is what the user will drag and ultimately drop
The user can drop the item onto another ScatterViewItem (placed in a separate ScatterView)
The overall interaction is quite similar to the ShoppingCart application provided in the Surface SDK, except that the source objects are contained within a ScatterView rather than a ListBox.
I'm unsure how to proceeded in order to enable the proper communication between my ViewModels in order to provide this functionality. The main issue I've encountered is replicating the ScatterViewItem when the user contacts the FrameworkElement.
You could use an attached property. Create an attached property and in the setproperty method bind to the droped event :
public static void SetDropCommand(ListView source, ICommand command)
{
source.Drop += (sender, args) =>
{
var data = args.Data.GetData("FileDrop");
command.Execute(data);
};
}
Then you can bind a command in your view model to the relevant control on the view. Obviously you may want to make your attached property apply to your specific control type rather than a listview.
Hope that helps.
I had a go at getting Steve Psaltis's idea working. It took a while - custom dependency properties are easy to get wrong. It looks to me like the SetXXX is the wrong place to put your side-effects - WPF doesn't have to go though there, it can go directly to DependencyObject.SetValue, but the PropertyChangedCallback will always be called.
So, here's complete and working the code for the custom attached property:
using System.Windows;
using System.Windows.Input;
namespace WpfApplication1
{
public static class PropertyHelper
{
public static readonly DependencyProperty DropCommandProperty = DependencyProperty.RegisterAttached(
"DropCommand",
typeof(ICommand),
typeof(PropertyHelper),
new PropertyMetadata(null, OnDropCommandChange));
public static void SetDropCommand(DependencyObject source, ICommand value)
{
source.SetValue(DropCommandProperty, value);
}
public static ICommand GetDropCommand(DependencyObject source)
{
return (ICommand)source.GetValue(DropCommandProperty);
}
private static void OnDropCommandChange(DependencyObject d, DependencyPropertyChangedEventArgs e)
{
ICommand command = e.NewValue as ICommand;
UIElement uiElement = d as UIElement;
if (command != null && uiElement != null)
{
uiElement.Drop += (sender, args) => command.Execute(args.Data);
}
// todo: if e.OldValue is not null, detatch the handler that references it
}
}
}
In the XAML markup where you want to use this, you can do e.g.
xmlns:local="clr-namespace:WpfApplication1"
...
<Button Content="Drop here" Padding="12" AllowDrop="True"
local:PropertyHelper.DropCommand="{Binding DropCommand}" />
.. And the rest is just making sure that your ViewModel, bindings and command is right.
This version passes an IDataObject through to the command which seems better to me - you can query it for files or whatever in the command. But that's just a current preference, not an essential feature of the answer.
I'm trying to implement a WPF application using MVVM (Model-View-ViewModel) pattern and I'd like to have the View part in a separate assembly (an EXE) from the Model and ViewModel parts (a DLL).
The twist here is to keep the Model/ViewModel assembly clear of any WPF dependency. The reason for this is I'd like to reuse it from executables with different (non-WPF) UI techs, for example WinForms or GTK# under Mono.
By default, this can't be done, because ViewModel exposes one or more ICommands. But the ICommand type is defined in the System.Windows.Input namespace, which belongs to the WPF!
So, is there a way to satisfy the WPF binding mechanism without using ICommand?
Thanks!
You should be able to define a single WPF custom routed command in your wpf layer and a single command handler class. All your WPF classes can bind to this one command with appropriate parameters.
The handler class can then translate the command to your own custom command interface that you define yourself in your ViewModel layer and is independent of WPF.
The simplest example would be a wrapper to a void delegate with an Execute method.
All you different GUI layers simply need to translate from their native command types to your custom command types in one location.
WinForms doesn't have the rich data binding and commands infrastructure needed to use a MVVM style view model.
Just like you can't reuse a web application MVC controllers in a client application (at least not without creating mountains of wrappers and adapters that in the end just make it harder to write and debug code without providing any value to the customer) you can't reuse a WPF MVVM in a WinForms application.
I haven't used GTK# on a real project so I have no idea what it can or can't do but I suspect MVVM isn't the optimal approach for GTK# anyway.
Try to move as much of the behavior of the application into the model, have a view model that only exposes data from the model and calls into the model based on commands with no logic in the view model.
Then for WinForms just remove the view model and call the model from the UI directly, or create another intermediate layer that is based on WinForms more limited data binding support.
Repeat for GTK# or write MVC controllers and views to give the model a web front-end.
Don't try to force one technology into a usage pattern that is optimized for another, don't write your own commands infrastructure from scratch (I've done it before, not my most productive choice), use the best tools for each technology.
Sorry Dave but I didn't like your solution very much. Firstly you have to code the plumbing for each command manually in code, then you have to configure the CommandRouter to know about each view/viewmodel association in the application.
I took a different approach.
I have an Mvvm utility assembly (which has no WPF dependencies) and which I use in my viewmodel. In that assembly I declare a custom ICommand interface, and a DelegateCommand class that implements that interface.
namespace CommonUtil.Mvvm
{
using System;
public interface ICommand
{
void Execute(object parameter);
bool CanExecute(object parameter);
event EventHandler CanExecuteChanged;
}
public class DelegateCommand : ICommand
{
public DelegateCommand(Action<object> execute) : this(execute, null)
{
}
public DelegateCommand(Action<object> execute, Func<object, bool> canExecute)
{
_execute = execute;
_canExecute = canExecute;
}
public void Execute(object parameter)
{
_execute(parameter);
}
public bool CanExecute(object parameter)
{
return _canExecute == null || _canExecute(parameter);
}
public event EventHandler CanExecuteChanged;
private readonly Action<object> _execute;
private readonly Func<object, bool> _canExecute;
}
}
I also have a Wpf library assembly (which does reference the System WPF libraries), which I reference from my WPF UI project. In that assembly I declare a CommandWrapper class which has the standard System.Windows.Input.ICommand interface. CommandWrapper is constructed using an instance of my custom ICommand and simply delegates Execute, CanExecute and CanExecuteChanged directly to my custom ICommand type.
namespace WpfUtil
{
using System;
using System.Windows.Input;
public class CommandWrapper : ICommand
{
// Public.
public CommandWrapper(CommonUtil.Mvvm.ICommand source)
{
_source = source;
_source.CanExecuteChanged += OnSource_CanExecuteChanged;
CommandManager.RequerySuggested += OnCommandManager_RequerySuggested;
}
public void Execute(object parameter)
{
_source.Execute(parameter);
}
public bool CanExecute(object parameter)
{
return _source.CanExecute(parameter);
}
public event System.EventHandler CanExecuteChanged = delegate { };
// Implementation.
private void OnSource_CanExecuteChanged(object sender, EventArgs args)
{
CanExecuteChanged(sender, args);
}
private void OnCommandManager_RequerySuggested(object sender, EventArgs args)
{
CanExecuteChanged(sender, args);
}
private readonly CommonUtil.Mvvm.ICommand _source;
}
}
In my Wpf assembly I also create a ValueConverter that when passed an instance of my custom ICommand spits out an instance of the Windows.Input.ICommand compatible CommandWrapper.
namespace WpfUtil
{
using System;
using System.Globalization;
using System.Windows.Data;
public class CommandConverter : IValueConverter
{
public object Convert(object value, Type targetType, object parameter, CultureInfo culture)
{
return new CommandWrapper((CommonUtil.Mvvm.ICommand)value);
}
public object ConvertBack(object value, Type targetType, object parameter, CultureInfo culture)
{
throw new System.NotImplementedException();
}
}
}
Now my viewmodels can expose commands as instances of my custom command type without having to have any dependency on WPF, and my UI can bind Windows.Input.ICommand commands to those viewmodels using my ValueConverter like so. (XAML namespace spam ommited).
<Window x:Class="Project1.MainWindow">
<Window.Resources>
<wpf:CommandConverter x:Key="_commandConv"/>
</Window.Resources>
<Grid>
<Button Content="Button1" Command="{Binding CustomCommandOnViewModel,
Converter={StaticResource _commandConv}}"/>
</Grid>
</Window>
Now if I'm really lazy (which I am) and can't be bothered to have to manually apply the CommandConverter every time then in my Wpf assembly I can create my own Binding subclass like this:
namespace WpfUtil
{
using System.Windows.Data;
public class CommandBindingExtension : Binding
{
public CommandBindingExtension(string path) : base(path)
{
Converter = new CommandConverter();
}
}
}
So now I can bind to my custom command type even more simply like so:
<Window x:Class="Project1.MainWindow"
xmlns:wpf="clr-namespace:WpfUtil;assembly=WpfUtil">
<Window.Resources>
<wpf:CommandConverter x:Key="_commandConv"/>
</Window.Resources>
<Grid>
<Button Content="Button1" Command="{wpf:CommandBinding CustomCommandOnViewModel}"/>
</Grid>
</Window>
I needed an example of this so I wrote one using various techniques.
I had a few design goals in mind
1 - keep it simple
2 - absolutely no code-behind in the view (Window class)
3 - demonstrate a dependency of only the System reference in the ViewModel class library.
4 - keep the business logic in the ViewModel and route directly to the appropriate methods without writing a bunch of "stub" methods.
Here's the code...
App.xaml (no StartupUri is the only thing worth noting)
<Application
x:Class="WpfApplicationCleanSeparation.App"
xmlns="http://schemas.microsoft.com/winfx/2006/xaml/presentation"
xmlns:x="http://schemas.microsoft.com/winfx/2006/xaml">
</Application>
App.xaml.cs (load up the main view)
using System.Windows;
using WpfApplicationCleanSeparation.ViewModels;
namespace WpfApplicationCleanSeparation
{
public partial class App
{
protected override void OnStartup(StartupEventArgs e)
{
var view = new MainView();
var viewModel = new MainViewModel();
view.InitializeComponent();
view.DataContext = viewModel;
CommandRouter.WireMainView(view, viewModel);
view.Show();
}
}
}
CommandRouter.cs (the magic)
using System.Windows.Input;
using WpfApplicationCleanSeparation.ViewModels;
namespace WpfApplicationCleanSeparation
{
public static class CommandRouter
{
static CommandRouter()
{
IncrementCounter = new RoutedCommand();
DecrementCounter = new RoutedCommand();
}
public static RoutedCommand IncrementCounter { get; private set; }
public static RoutedCommand DecrementCounter { get; private set; }
public static void WireMainView(MainView view, MainViewModel viewModel)
{
if (view == null || viewModel == null) return;
view.CommandBindings.Add(
new CommandBinding(
IncrementCounter,
(位1, 位2) => viewModel.IncrementCounter(),
(位1, 位2) =>
{
位2.CanExecute = true;
位2.Handled = true;
}));
view.CommandBindings.Add(
new CommandBinding(
DecrementCounter,
(位1, 位2) => viewModel.DecrementCounter(),
(位1, 位2) =>
{
位2.CanExecute = true;
位2.Handled = true;
}));
}
}
}
MainView.xaml (there is NO code-behind, literally deleted!)
<Window
x:Class="WpfApplicationCleanSeparation.MainView"
xmlns="http://schemas.microsoft.com/winfx/2006/xaml/presentation"
xmlns:x="http://schemas.microsoft.com/winfx/2006/xaml"
xmlns:WpfApplicationCleanSeparation="clr-namespace:WpfApplicationCleanSeparation"
Title="MainWindow"
Height="100"
Width="100">
<StackPanel>
<TextBlock Text="{Binding Counter}"></TextBlock>
<Button Content="Decrement" Command="WpfApplicationCleanSeparation:CommandRouter.DecrementCounter"></Button>
<Button Content="Increment" Command="WpfApplicationCleanSeparation:CommandRouter.IncrementCounter"></Button>
</StackPanel>
</Window>
MainViewModel.cs (includes the actual Model as well since this example is so simplified, please excuse the derailing of the MVVM pattern.
using System.ComponentModel;
namespace WpfApplicationCleanSeparation.ViewModels
{
public class CounterModel
{
public int Data { get; private set; }
public void IncrementCounter()
{
Data++;
}
public void DecrementCounter()
{
Data--;
}
}
public class MainViewModel : INotifyPropertyChanged
{
private CounterModel Model { get; set; }
public event PropertyChangedEventHandler PropertyChanged = delegate { };
public MainViewModel()
{
Model = new CounterModel();
}
public int Counter
{
get { return Model.Data; }
}
public void IncrementCounter()
{
Model.IncrementCounter();
PropertyChanged(this, new PropertyChangedEventArgs("Counter"));
}
public void DecrementCounter()
{
Model.DecrementCounter();
PropertyChanged(this, new PropertyChangedEventArgs("Counter"));
}
}
}
Just a quick and dirty and I hope it's useful to someone. I saw a few different approaches through various Google's but nothing was quite as simple and easy to implement with the least amount of code possible that I wanted. If there's a way to simplify even further please let me know, thanks.
Happy Coding :)
EDIT: To simplify my own code, you might find this useful for making the Adds into one-liners.
private static void Wire(this UIElement element, RoutedCommand command, Action action)
{
element.CommandBindings.Add(new CommandBinding(command, (sender, e) => action(), (sender, e) => { e.CanExecute = true; }));
}
Instead of the VM exposing commands, just expose methods. Then use attached behaviors to bind events to the methods, or if you need a command, use an ICommand that can delegate to these methods and create the command through attached behaviors.
Off course this is possible. You can create just another level of abstraction.
Add you own IMyCommand interface similar or same as ICommand and use that.
Take a look at my current MVVM solution that solves most of the issues you mentioned yet its completely abstracted from platform specific things and can be reused. Also i used no code-behind only binding with DelegateCommands that implement ICommand. Dialog is basically a View - a separate control that has its own ViewModel and it is shown from the ViewModel of the main screen but triggered from the UI via DelagateCommand binding.
See full Silverlight 4 solution here Modal dialogs with MVVM and Silverlight 4
I think you are separating your Project at wrong point. I think you should share your model and business logic classes only.
VM is an adaptation of model to suit WPF Views. I would keep VM simple and do just that.
I can't imagine forcing MVVM upon Winforms. OTOH having just model & bussiness logic, you can inject those directly into a Form if needed.
" you can't reuse a WPF MVVM in a WinForms application"
For this please see url http://waf.codeplex.com/ , i have used MVVM in Win Form, now whenver i would like to upgrade application's presentation from Win Form to WPF, it will be changed with no change in application logic,
But i have one issue with reusing ViewModel in Asp.net MVC, so i can make same Desktop win application in Web without or less change in Application logic..
Thanks...