Entries Tagged 'Learning Silverlight' ↓

Shawn Wildermuth, theADOguy (who should really rename himself the Silverlight guy) is giving a free seminar at the Microsoft office in Dallas on 2/18. Shawn is a very smart cool guy who knows a lot about a lot, and even more about Silverlight, so this should be a great event, and I'm told he's capping it at 32 seats so "don't say I didn't warn yer."

Pragmatic ADO.NET: Data Access for the Internet World (Microsoft .NET Development Series) by Shawn Wildermuth Read more about this title...

Today's Webcast focused on factoring out repeated creation of objects in XAML (and their manipulation in the code behind) by creating objects dynamically using CreateFromXaml (This will be available for viewing as soon as the recording is processed).

The webcast began by showing this application, that uses animation to display changes in data over time.

It then explored the problems with this approach and offered a solution as shown in this video.

During the course of the webcast I mentioned a number of postings that you may want to read:

• A tip of the day covering CreateFromXAML
• Tip of the Day showing that you can retrieve the Silverlight control (plugin) using GetHost()
• Brief Tip of the Day pointing to information about the Silverlight control's content sub-object
• Introductory Video on using Silverlight with Web Services
• More advanced video that uses Silverlight with Web Services, AJAX and JSON

Thanks again.

Do not miss Adam Kinney's excellent tutorial on creating a bar chart in Silverlight 1.0; it is brilliant, and it is a great staring point on working with XAML, JavaScript, JSON data and many other interesting aspects of Silverlight.

Also, while I have you, a quick apology; in my haste to put up today's Tip of the Day I posted a msg. that I would be late; brainlessly forgetting that said message would be sent out via RSS wasting everyone's time and bits. Won't happen again.

I have had a very strong positive reaction to focusing on the Javascript in code behind for Silverlight 1.0, and so will devote not only today's Tip of the Day to a bit more on the topic, but will make sure that I do so on a frequent basis for a while.

Before I start: a small house keeping measure: the convention of starting every Tip of the Day with "Did You Know" is getting old. Thus, starting with the very next one it will be replaced with Tip Of The Day:

A reader (let's call this one Joe) writes,

What I've had trouble understanding is a small bit of code....

 Now, I have to say that the best way to get this kind of question answered is to post it to our forums; that way a lot more people see it a lot more quickly; but since you did write to me, I'll feel free to use it as grist for the Tip of the Day mill...

sender.fill = createLinearGradientBrush(sender.getHost(), sender);

function createLinearGradientBrush(plugin)
{
    var xamlFragment = '<RadialGradientBrush>';
    xamlFragment += '<GradientStop Color="LimeGreen" Offset="0.0" />';
    xamlFragment += '<GradientStop Color="Green" Offset="1.0" />';
    xamlFragment += '</RadialGradientBrush>';

    return plugin.content.createFromXaml(xamlFragment, false);
}

I understand 'what' it is doing and i understand 'why' it is doing it, but I don't know 'how'... I found the code to do it on a forum return plugin.content.createFromXaml(xamlFragment, false)

It isn't surprising you're baffled by this code as it is pretty advanced and it starts in the middle.

sender.fill = createLinearGradientBrush(sender.getHost(), sender);

What can we see or guess?

1. the purpose of this line is to fill sender with a LinearGradientBrush.

2.  the LinearGradientBrush will be created by the function createLinearGradientBrush.

3. The function createLinearGradientBrush takes two parameters: the getHost() method of the sender and the sender itself

4. The function crateLinearGradientBrush returns a Brush (and it better be a LinearGradientBrush or someone should be hurt) that be assigned to the sender's fill property.

What is sender?  It's not a big leap to assume sender is a shape.

How did we get here? We assume "here" is the middle of some function, and sender was a parameter to that function.

To do this right, I think we want to create a simple application that might get us to just this point.  So I'll fire up Visual Studio, strip out the starter application and put the following into Scene.xaml:

This creates two shapes, both of which respond to the MouseLeftButtonUp by calling an event handler method named onMouseUp, which we will expect to find in the code-behind file.

Note that you must provide a fill or if you click inside the shapes the event handlers will not fire! If you want the boxes to appear unfilled, you can use Fill="Transparent" as I've done here, but if you leave the fill out, it will default to null and clicking inside the shape will have no effect.

Here's the event handler:

  Presumably the event handler has work to do or there is more than one event handler. In any case, the programmer has decided to factor out the work of dynamically creating the Linear Gradient Brush. We now know what sender is; it is whatever shape we clicked on (in this case) but of course you can imagine other programs in which sender will be a shape retrieved by other means (picked from a menu in a drawing program?)

We saw CreateLinearGradientBrush in the original code above, but it is needlessly complex and we can simplify both the call to it and its own logic. Let's send just the sender, and encapsulate within createLinearGradientBrush the work of getting the SilverligthHost. First, we change the call,

Notice that we no longer ask the calling method to know that we need to retrieve the Silverlight control by calling GetHost; we just pass along the sender (the shape). Next, we create the brush by creating the XAML as a string and then asking the Silverlight control to create the brush we need from that string,

The top half of the method creates a string that replicates exactly how you would create the brush in XAML. To make it clear, I've put in the spacing I'd use in a XMAL file. Though this is not required, I believe it is good coding practice and far easier to maintain.

The bottom three lines (above the comment) act as follows;

var SilverligthHost = sender.GetHost()

GetHost() can be called by any Silverlight object as explained in this article and in the Silverlight 1.0 Help Files (an often overlooked and quite useful resource

Now, I admit it takes a while to translate some of that back into English, but what it tells you is that you can call GetHost on any Silverlight object (e.g., sender) and get back the plug-in (that is, the Silverlight control). Handy.

The second line is the key to this entire exercise

var brush = silverlightHost.content.createFromXAML(xamlFragment, false);

To understand this, you need to know that there is a method, createFromXAML that takes a string of XAML and returns a XAML object. That is explained in some detail in this video.

In addition it is helpful to know that the Silverlight control has sub-objects, one of which is named content, as explained in

You also want to know that createFromXAML takes two arguments, the second of which is used to create a unique namespace for the objects it creates, as explained both in the video and in the Silverlight documentation

I've commented out the line that is actually the way most programmers (or at least most of us who grew up in the C/Unix world where terseness is prized despite the fact that it makes for harder to maintain or understand code, and we should all be slapped silly)

return sender.GetHost().content.createFromXaml(xamlFragment, false);

I include it here not because I think it is the right way to do it, but because you'll see it and will want to know how to read it; the answer is "inside out" --

"Return from this method the result of calling GetHost on the sender object and then accessing the content sub-object on the returned Silverlight control so that you can call createFromXAML and pass in the text fragment that represents the object you are creating plus the boolean value false signifying that you do not want to protect namespaces." 

Easy.

So; when you bring up the application, the objects look like they are unfilled (their fill is transparent, but if you click on them, the event handler is called, which in turn calls createLinearGradientBrush passing in a reference to whichever object was clicked on. That object accesses the Silverlight control and asks it to instantiate a brush, which it uses to fill itself with color,

Hope that helps.

-j

A very insightful reader sent me email that I'm going to respond to here. To protect his identity let's call him Bob. I'm going to excerpt Bob's  email and rephrase a bit so that this becomes a more general discussion.

I find that a lot of Webcasts and books about Silverlight spend most of their time explaining XAML or using Expression tools in  great depth, yet whenever its time to examine the code,  the presenter tends to speed up and glide over the actual JavaScript.

...I noticed in your last Webcast you made the comment that you would be spending most of the hour talking code, ...however, once again lots of time was spent on XAML shapes and gradiant fills, but when it came to the actual code behind side, that was rapidly glossed over....I get the feeling that a lot of people are really enthusiastic about Silverlight, however they are perhaps lacking the "JavaScript 101 for Silverlight" that is needed to follow along with most of the webcasts.

Bob went on to observe that switching rapidly between Visual Studio and IE in Live Meeting is not a great tactic (one I will solve next time by setting up my desktop properly, and then sharing the entire desktop to avoid all the switching).

I think Bob's key point is exactly right, and while I can't answer for anyone else, I can tell you why this happens in my presentations; though now that it has been pointed out to me, I have the opportunity to make a conscious decision to change.

First, in a course named "Getting started with..." or "from scratch" there seems to be an imperative to spend a good bit of time showing where the resources are. That can be very time consuming. It might be better to point folks to our web site and to my How Do I Get Started? web post and let it go at that.

The second reason is more pressing, and it is that Silverlight 1.0 has  always been about XAML and code-behind, but the code behind has always been, if I'm going to be totally honest, "well, it's JavaScript right now, but it should be C# and it will be C#, so let's just show how the logic works without dwelling on the implementation which is, after all, not object oriented or the way we'll do things in the long run."

Now, there are a couple real problems with that attitude. 

First, Silverlight 2 won't be released (RTM) for quite a while. We'll be in Beta with a Go-Live license very soon, but it won't be replacing 1.0 until it is released. We've not announced a release date. Which means 1.0 isn't going away next week. So glossing over coding in 1.0 is a mistake.

Second, you can still code in JavaScript if you want to in 2.0 and some folks will most certainly want to (Calling all AJAX programmers).

Most important, if you are going to teach Silverlight From Scratch, then you can't gloss over the implementation of the code behind (read, in most cases, the event handlers) because you run the risk (read certainty) of leaving your users hopelessly confused and frustrated.  It all becomes a hand-wave, and even worse, the critical distinction between the XAML file and its code-behind becomes even more confusing. So I must stop this madness.

Hooking Up the Event Handlers

In the figure, Scene.xaml declares a Canvas and declares that the Canvas (which will hold a diamond) will respond to three known events: MouseLeftButtonDown, MouseLeftButtonUp, and MouseMove. The three functions that will serve as event handlers are named: OnMouseDown, OnMouseUp and OnMouseMove respectively.  

These Event handlers do not exist in the XAML file. They exist in the code-behind file: Scene.xaml.js

Note that the programmer could have named the functions Moe Larry and Curly; as long as the names match in the XAML declaration and the code-behind file, all is good.

This is where I wave my hands at the JavaScript and move on. So, why? 

JavaScript isn't that hard, though I will tell you that you'll run into two basic approaches when you review example code here and elsewhere. One, as shown here is the older style of independent functions and global variables. Perfectly respectable. The other, more object-oriented-ish creates kinda-sorta-member methods with kinda-sorta-member variables. That looks more like this:

In this example (which is what Visual Studio 2008 creates as a sample for you when you make a new Silverlight app) the prototype acts as a "class" and each function acts as a "method."  Any variable prepended with "this" is a member, declared with var is local.  In this style it is more common to wire up even handlers using addEventListener in the code rather than in the XAML.  More about this approach here.

It's Just Code

Returning to our story, already in progress, we had three event handlers: onMouseDown, onMouseMove and onMouseUp.

Each of these is just JavaScript, the kind you can find incredibly well described here in David Flanagan's book, BLOCKED SCRIPT The Definitive Guide:

JavaScript The Definitive Guide by David Flanagan Read more about this title...

The code behind begins by declaring three variables that will be used by the event handlers.  The job of OnMouseDown is to store the location of the mouse by calling GetPosition on the mouseEventArgs argument passed in with the syntax shown. We then set trackingMouseMove to true (to signal to OnMouseMove that we are tracking a moving object) and we tell the sender (the object that caused the event) to "capture" the mouse, so that if we move off that object we'll still get mouse move events until the user lets the mouse button up.

As the user moves the mouse we check to see if we are tracking. If so, we continually recompute the new position of the mouse and then recompute the new position of the object, setting its Canvas.Left and Canvas.Top properties and then resetting the new values for the global  beginX and beginY

Finally, when the user lets the left mouse button up, we stop tracking and we tell the object to "uncapture" the mouse.

Not so bad.  Other than the manipulation of currentX vs BeginX in MouseMove the logic is fairly intuitive, and the syntax should feel pretty familiar to anyone who has worked in the C/Java family of languages.

All of this is walked through in great detail in this video and its example code.

Accessing XMAL defined objects in Code

Since I'm up late, and on a roll, let me mention one more bit about the relationship between XAML and the code behind; and that is how you access an object declared in XAML from your Javascript. The key to making this work is

1. Name the object in XAML
2. Use FindName to assign the object to a variable in the code behind

The XAML file:

Referring to these objects in code can be done in a number of ways, but it typically involves calling FindName, passing in the name used in the XAML as a string, and returning the reference to an object in the code:

Note that plugIn is the Silverlight Control which is passed in as the first parameter to hanldeLoad.  If you are not in handleLoad you can always get a reference to the Silverlight control by calling GetHost() as explained here.

Of course there's more to say, but this is a start on addressing the well made point that understanding the Javascript is an essential point in understanding how the Silverlight application works its magic.

Much thanks to "Bob"

I will be giving my last pre-mix Webcast on Feb 6 (did i hear cheering??) on Feb 6 on the topic: "Building a dynamic Silverlight Application using a web service, the downloader and CreateFromXAML" -- you can read more about it and register for it through the MSDN Events Site

I have a series of Webcasts planned about Silverlight 2 once it is released, and I will post that schedule as soon as (a) we unveil <smile> and (b) I learn how to post to the Microsoft events site (which you will not be shocked to learn is non-trivial)

JFK (the O'Bama of my childhood) is reported to have said about Washington DC that it "combined Northern charm with Southern efficiency. "

I find that Webcasts combine the demand for all the organization and flow of ideas of a live presentation without the feedback of having an audience in the room with you. That is, they're hard to do well. Add to that my total inability to manage the microphone, keyboard and windows at the same time and... well bear with me... the material will be great.

In fact, this last 1.0 web cast (for a  while) should be pretty interesting because we'll be looking at how you can add to your Silverlight control at run time based on XAML that you obtain from a web service.

One final note; after 2.0 is released, we'll have 1.0 out as a released product, and 2.0 as a Beta product. I personally will not be walking away from 1.0 at all; though of course I'll be spending a lot of time looking at 2.0 (everyone loves the puppy, but never ignore the dog).

Technorati Profile

Thank you if you were able to join our Webcast Silverlight For Total Novices. (If you were unable to attend, please keep an eye on the events page for the Recording.)

In the time we had, we could only scratch the surface, but a number of resources were mentioned and I wanted to make sure that I posted links to them so that you can find them easily. While I'm at it I'll add a few other valuable resoruces that were not mentioned.

In imperfect order they include:

1. The best starting point is the Silverlight Site

2. You may want to read my post Silverlight For Total Novices that reviews the steps for downloading what you need and recommends which videos to watch and in what order (This post by Tim Heuer explains why you need the Silverlight tools for Visual Studio 2008 even though you are building 1.0 applications and the tool says it is for 1.1!)

3. A great next stop is the Fire Starter tutorials.

4. The "How Do I" videos on creating shapes, fill, transforms, drag and drop and streaming are all here.

5. If you are interested in what is coming in Silverlight 2.0, Scott Guthrie's announcement is the third item in this blog entry. 

6. Questions and Answers can be found on our forum

Thanks again and best of luck!

Technorati Profile

Adam Kinney has worked with Dave Campbell to feed Silverlight Cream to Twitter.

You can subscribe here. In addition, Adam has created a script for placing this news into his (or your!) web log and formatting it.

You can download the script here, along and read full details on how to use it on your own site as well. Once in place, its all automagic.

More on this to come!

Technorati Profile