Street Fighter Style UI Compositing With PrismSubmitted by smartyP on Tue, 01/05/2010 - 21:29
If you're into Silverlight or WPF, then you've undoubtedly heard the nonstop clamoring over Composite Application Guidance for WPF and Silverlight - aka Prism. I have a list of things in Evernote that I'd love to write about Prism, but I figured I would write about something I actually felt like typing up instead - something I tend to call Street Fighter style compositing in Prism.
Recap on UI Compositing in Prism
You already know one of the bigger parts of Prism is the UI compositing it supports. Regions are defined in the Shell.xaml and you inject modules into them. The idea being that you can add, remove, and swap out sections of your application dynamically. It's one of many great technologies in Prism.
Street Fighter Style Compositing in Prism
Having done a few Prism applications, I've found that the composite UI support in Prism is really snazzy, but I've also found that I don't tend to use it the way that most Prism examples show it off. While I will frequently define region areas like shown in the example above, I will just as frequently use regions for layering - kinda like how they did it in Street Fighter.
If you've played Street Fighter before then surely you've noticed that the UI is made up of layers on top of layers. The bottom layer is the static stage background, a layer on top of that might be the animated characters in the background, the layer above that may be where the fighters are shown, above that a layer for projectiles, above that a layer for health bars and high scores, and a layer above that for the menu - which is currently hidden.
If you're going to try and do this kind of layering with regions in Prism then one thing to keep handy is how to create a Prism region whose contents will not specify a specific size, but will scale to fit the entire region. You'll want this code snippet for such a region:
<!-- image popup area -->
The reason I find myself using this style of composition is that it allows a lot more flexibility if you're not creating a rigid desktop style layout. Using full window overlays will make it easy to place modules into the region centered horizontally and vertically - popup menus or loading status windows for example. There are some really neat things you can do with this kind of layering, especially with perceived depth - but that's not really the point of this article.
A Quick Example
Regions used in ImageWind.net alongside its current appearance
To use Image Wind as an example, here is a rough outline of the regions used (from back to front):
- background layer
- image side scrolling layer
- title region (upper left)
- controller region (lower left)
- full size image display popup layer
- popup dialog / loading layer (various modules displayed in center)
- beta badge and service error region (upper right)
The main point of this article wasn't to show the cool stuff you can do with layering, or to cover when you might want to do layering within the same region versus in different regions, etc. The point of this article was really just to make sure that the composite UI in Prism isn't always thought of in such a rigid way. Oh, and to coin the term 'Street Fighter' style compositing.
A rough outline of the region layout from another recent project
Hi there, this is a very
Hi there, this is a very interesting article that is really out of the box kind of thinking. I don't suppose you have source you'd be willing to share so I can see how this works?
What I'm most interested in is to see how you tell controls/views to load/unload themselves from the parent view's ItemsControl. For example, are you subscribing to events on the user controls and unloading them from the parent's ItemsControl when a CloseEvent (or similar) is fired?
I'm also very curious to see how you handle allowing of interaction with other layered usercontrols. For example, in your picture, i see a listbox and a cloud control that appears to sit on top of it. Can you actually click on the listbox and interact with it anywhere between the cloud's crevices? Does that make sense? It'd almost have to work like an html map to get it to work like that. Just curious as to how you handle that if that's indeed how it functions.