All posts by admin

Microsoft Band API

So I just started looking at the SDK for the Microsoft Band. It’s pretty good, but bear in mind that there are limitations, for example, you can only gather “live” sensor data – there’s no way to get any cached stats that may (or may not) be on your device.

Having said that, here’s a quick overview of using the SDK.

Microsoft Band
The Band

Preparing the solution

First thing to do here is to make sure your band is paired with your phone, turned on and active. Once that’s done, open Visual Studio and create a new phone project.

The SDK is available via NuGet, so add a reference to it.

You should see that references have been added to the required assemnblies, but also your package.appxmanifest file will have been updated with the following device capabilities.

These allow the app to communicate with your Band.

Now you can get going.

Connecting to your Band

In your page, get a list of the paired band devices on the system with the following command. Note that I return from the method if there are no paired Bands. You’ll need to set a using reference to Microsoft.Band.

Once you have your list of Bands, you can connect to one. In this case, I’m assuming we always want to connect to the first connected device, but you can modify this bit to suit your needs.

Note: The tutorials suggest you wrap the returned client instance in a using block, but if you do that then later steps like collecting sensor data won’t work because the client is disposed when this method returns. I hold a local reference to the client object and dispose it in OnNavigatedFrom();

At this point we can start getting basic data back, like the hardware and software versions.

Gathering sensor data

Now we get to the fun stuff. Assuming you followed my advice above about holding a local reference to the client object, you can start getting sensor data back. In this case, the HeartRate data.

Add a using statement for Microsoft,.Band.Sensors, then you can subscribe to events.

Since all this work takes place on its own thread, you’ll need to do a little bit of juggling to get this data onto the screen. Specifically use the CoreDispatcher to invoke the action on the UI thread. You might like to set up a method to keep this simple…

and then call it like so…

for example.

Next time – more on sending notifications and personalising your Band.

Accessibility – upcoming series

Although I’m currently working on a large project, I can’t stop thinking about the state of accessibility in apps available on the Windows Stores.

I am lucky enough to be in a position where the accepted use of the term “accessibility” doesn’t apply to me*, but I do wonder how many apps out there have given any thought to the topic of enabling their application for use by people with disabilities or other factors affecting their use of computers.

The Windows Store frameworks have a lot of inbuilt functionality to support users with enhanced requirements from their apps (documented here), but how many of us, as developers, have put in the time and effort to make sure we can reach the widest possible audience?

I am hoping to start work soon an a series of how-to posts on making the most of the platform’s feature to reach all our potential users.  In the meantime, if you have any specific accessibility questions you would like to see me address, please drop me a line.

* although accessibility applies to everyone, to varying degrees. We might not be able to read AntiqueWhite text on a white background, and no designer in their right minds would expect us to. I’m more interested, for the purposes of this post, in enabling apps for those with visual impairments or motor control issues.

Nokia Imaging SDK updated

Not surprising news, but worthy of a big mention. The Nokia Imaging SDK has had a new beta release (1.2).

This release brings all the goodness of the Phone library to the new, conjoined Microsoft platform across Windows 8.1 and Windows Phone 8.1. I am hoping to find a little time to play with the new SDK over the next week or so, but the sample apps and code look great. There are also sample apps in the Windows store, although the links in the site seem not to work. I’d suggest opening the Windows Store and searching for “Filter Effects” and installing the one published by Nokia.

It’s //build time!

For those of you who haven’t managed to convince your management to let you go to //build (and this includes me), you can see what you’re missing on the official //build 2014 site. The recordings of the sessions will also be made available there.


Having had a look at the session listings, there is big (although maybe not unexpected) news for Windows Phone and Windows Store apps. As ever, you can find the details 24-48 hours after the live events in the videos and slide decks (or just keep an eye on your favourite Twitter client / blog).

Windows Store Apps with Prism and Unity

It’s been mighty quiet around here. Sorry about that, I’ve been up to my neck in a new project that’s been keeping me pretty busy.

I am going to start a mini-series on developing Windows Store apps using Prism and Unity.

The first of the series (in other words, this post) will look at how to separate your UI and ViewModels into separate projects, but more specifically, how to override the default behaviour to look up the view models from the appropriate place.

By default, Prism will look for your views in the “Views” folder, and take the name of the view (e.g. MainPage.xaml) and use that to look for the ViewModel in the same project’s ViewModels folder. The basics of what it does are to remove the “Page” part from the end of the class name, replace it with ViewModel and look up the class in the x.ViewModels namespace. Of course, if you want to put your ViewModels in another project for maintainability purposes, you’ll need to do a bit of dancing to get it to work.

Thankfully, it’s pretty simple. In your main UI project’s App.xaml.cs, locate the OnInitialise method and add the following code.

Basically, we get the fully qualified name of the target assembly and class type, turn that into a type, and then return it to the view model locator. In the example above, my main UI project is called “MyProject.UI”, and my UI logic (with the view models) is contained in “MyProject.UILogic”.

One thing to notice from my code is that it will also account for namespaced folders beneath the Views folder. The default example implementations don’t take this into account, so the view “Views\Flyouts\AboutPage.xaml” wouldn’t resolve the ViewModel if you follow the same folder structure in the ViewModels folder. Don’t worry, though – as long as you create the same sub-folder structure and have your ViewModels in the correct location, the code above will work.

Don’t forget, though, you’ll need to manage the strings in the snippet above so that they match your assemblies’ fully qualified names. I’ll leave that as an exercise for you, though.

Unit Testing in VS 2013

Visual Studio 2013 now contains functionality to write tests against a Windows Store project. Now, it is a relatively simple task to write unit tests against the code you have in classes within your app’s structure.

A Basic Example

Launch Visual Studio, and from the new project dialog, select Templates -> Visual C# -> Windows Store, and create a new “Blank App (XAML)” named “AppToTest”. Make sure you change the Solution Name value to be “TestableStoreAppExample”


Within the project, create a new folder named “DataAcccess”, and within that a new class named “SampleDataAccess.cs”

Open your newly created class, and mark it as public.

Now we want to add some methods to the class. These will be the methods against which our tests will run, so we will just create some dummy ones for now.

As you can see, the content of the class is very basic at this stage. We will build on this later.

Build the project to ensure that there are no errors.

Next, Let’s add a unit test project.

Right click on the solution in Solution explorer, and click on “Add -> New Project”. In the dialog, select “Unit Test Library (Windows Store Apps”, and name the test project “AppTests”.

In the unit test project you just created, right click on References and select “Add Reference”. In the dialog that appears, navigate to “Solution” on the left hand side, and select “AppToTest on the right, then press OK.

02 - AddReference

Now we can write some tests. For now, I’m just going to add these tests into the default class created when the project was initialised.

Open the file “UnitTest1.cs”

There are couple of annotations to note in this class, but these should be familiar to you if you have used MSTest before.

[TestClass] Marks the class as being a test fixture, containing unit test code.
[TestMethod] Marks a method as being a Unit Test.

For our first test, I’ll re-use the code already in the class. First of all, let’s change the name of TestMethod1 to something more descriptive. Since this test will call the GetSumOfInts method we defined earlier, lets call it TestIntegerSummation().

To break that down, first I create a new instance of the SampleDataAccess class, and call the GetSumOfInts method with 1 and 2, hopefully giving a result of 3. The final lines is where the test comparison happens. In this case, I’m calling Assert.AreEqual to test that the expected value of 3 matches the actual returned value. There are a few other Assert. methods available, e.g. Assert.AreNotEqual, Assert.IsTrue, etc.

Now build the solution in order to make the test available to the test engine.

To run the method, my preferred method is to open the Test menu in Visual Studio, and select Windows -> Test Explorer.


From here you can run all tests, or right click on a test name to run individual ones. (Note, theres another feature that you might want to take advantage of – automatically run tests when you build the solution. To do this, go to Test -> Test Settings and select “Run Tests After Build)

The screenshot above shows that the test we wrote passes.

OK, that’s a synchronous method. What about async ones?

Testing async Methods

This isn’t as intuitive as you might imagine, but once you know the trick, it’s very simple.

As you saw when we created the class to test, I created an async method that goes off and pulls text from a web resource. This is an async call due to the potential delays in loading a web resource.

To create a test for this, we first need to define a new method in the test class and decorate it with the [TestMethod] attribute, then write the code we want to call in the method.

If we do this with the same signature as before, however, the test won’t appear in the test explorer. There are two things we need to do to make the test runnable.

1) Add the async keyword to the method signature (to allow the use of the await keyword in the method call)

2) Change the return type of the test method from void to Task (in the System.Threading.Tasks namespace)

Our method should now look like this.

The test now shows up in Test Explorer and you can run it just like any other method.

More Complex tests

Another great way to test methods is by the use of the [DataTestMethod] attribute. This attribute allows you to specify many sets of data to run in the test, without the need to refactor a test to support different values.

Let’s take our TestIntegerSummation() test and refactor that to support different data sets.

Make a copy of the method, and call it TestIntegerSummationWithData, and add three integer parameters to the signature, calling them “input1”, “input2” and “expectedResult”.

Next we need to update the method body to use these values. The result should look like this.

Change the Attribute from [TestMethod] to [DataTestMethod].

To add sets of data against which we want the tests to run, we can add additional attributes, in this case, the [DataRow()] attribute.

Each DataRow attribute allows you to specify values to be used within the method (in the order they are defined in the method signature). The data we want to pass in should be a reasonable set of data that can be used to test various conditions, but for the purposes of this post, we’ll just pick some random values and test those.

You can see that there are likely to be some edge cases where the calculation will not return the expected result, for example, if we set input1 to be Int.MaxValue. In that case, running the test will show the following in Test Explorer – you can see it gives individual results for each DataRow defined.


There are a couple of ways to handle this situation. We can simple extend the method signature and data rows to pass in a value of whether we expect the test to pass or not

But this isn’t an ideal situation, especially if we expect an exception.

I’ll talk about better ways to handle this in the next post.

Remembering files from outside the application repository

A common scenario for file access in Windows 8 is to allow the user to pick a file from their hard drive, not necessarily from within one of the system-allowable locations such as the local data folder, pictures or documents library, etc. This is, of course, easily achievable using the FileOpenPicker control, but what happens if you want to persist that selection across runs of your app? If you just store a reference to the file location, subsequent runs won’t be granted access to the file and that could bring your app logic crashing down.

The answer is fairly simple. Use the FutureAccessList object in the StorageApplicationPermissions namespace. This allows you to store a file (or folder) along with a token in order to be able to use the IStorageItem later on.

This is a simple example page that uses the FutureAccessList to allow the user to set an image source, which is persisted across runs of the app. For the sake of clarity, I have removed XML comments from the code.



So, what is happening here?

1) In OnNavigatedTo, the app will check to see if we have already set an image file. Let’s assume for now that we haven’t and skip forward.
2) When the user clicks the “Load Image” button, we show a filtered FileOpenPicker to pick a single file. Once we have the file back, we stash it in the FutureAccessList, along with a token to identify the file – in this case, just a string “pickedFileToken”.

3) I have moved the logic to render the file content to the “MyImageWithFileAccess” image out to a separate method. This opens a stream and sets the image source accordingly.
4) On the next run of the app (or visit to the page), we go back in to OnNavigatedTo. This method checks the FutureAccessList to see if we have set an image. In this case, we have, so it retrieves the StorageFile instance from the list and calls the SetImagePath method once more.


  • There is no error handling here – if the user deletes the file, we will get an exception.
  • We are not removing items from the FutureAccessList if we no longer need them. This can be done by calling the Remove method on the list, passing in the token, or clearing the list entirely with the Clear() method.
  • The FutureAccessList can hold up to around 1000 items, but try to use it sparingly. If you are having to keep track of 1000 items, you may need to re-think your app logic.

Javascript Windows Apps

I’m a XAML / C# kind of guy. I haven’t delved very much into the JavaScript side of W8 app development, but when browsing the store earlier today I cam across the CodeShow() app. (store link)

CodeShow() Logo
CodeShow() Logo

If you are running with a JavaScript project, there are plenty of examples in the app of how to achieve certain things with working demos and source code. Seems like it could be a useful resource.

*Disclaimer* : I am not associated in any way with the author of this app – I’m just pointing to it in case someone finds it useful.

Discrete Slider in Windows Phone

This morning I found an interesting question on StackOverflow asking how to make the default Windows Phone slider control return discrete values (in this case an integer value rather than the default double).

There are several possible ways to go about this, but the one that leapt out was to create an attached property on the slider and bind to that.

The entire helper class for the property looks like this.

The points to note here are:

  1. On line 7 we hook into the slider’s ValueChanged event so that we can round the value.
  2. Ob line 12 I just return the slider’s Value property, cast as an integer. This value has already been rounded and set on line 7.
  3. If we don’t attach the property, the slider will continue to behave as normal

To use this code in an app, here’s a sample of three sliders, each with a TextBlock to display the Value.

I know this isn’t an elegant solution to the problem, and that it has a couple of issues, but it does work and should give the results the original questioner was after.

Syncfusion e-books

Syncfusion have released a series of e-books on various topics, including Windows Store Apps. You need to enter some registration information to be able to access them, but the content looks good at first glance – certainly good enough to make it worth registering.