MVVM, Josh Smith’s Way

I’ve long admired Josh Smith’s work. He’s one of the best explainers on the web and he – together with Karl Shifflett and Andrew Smith - gave us Mole, the superb WPF Visual Studio debugger visualizer.

I heard recently that he’d published an e-book on Model-View-ViewModel (MVVM). I met him for the first time at the Microsoft MVP Summit last week and he graciously handed me a printed copy. I devoured it on the plane home and spent the next day playing with the code and preparing this review.

It’s called “Advanced MVVM” and you can buy a printed copy for $20 or get it for $15 on the Kindle. The companion code is freely available at AdvancedMvvm.com.

$15 bucks for a 52 page “book”? It’s more of an extended essay to be honest … an essay with code. After spending some pleasurable hours with both book and code I thought “where do you get that kind of entertainment / self-reflection for $15?” I have a surfeit of 600 page sleepers lounging on my bookshelf at home; paid more then $50 for each and hardly cracked them. Should we pay by the pound? So I’m over it. I don’t mind paying for the hours Josh poured in and the fun I got out. Set your expectations appropriately and you’ll feel fine.

---- 19 March 2010 Note ---

Ok, it’s been a month or so since I wrote this. Life happens while you’re making plans. I intended to be among the first to comment on it. I’m late to that party (although I have deliberately NOT looked at what others may have said) so I’m not sure if it’s still timely. I also shared it with Josh last week to see if I had missed something and get his feedback, particularly in light of my critical remarks.

He encouraged me to publish it even at this late date. I hope he’ll repost here some of his responses to the points I raised.

----------------------------------

MVVM has spooked a lot of developers. There’s more than a whiff of astronaut architecture about it. There have to be at least 20 MVVM frameworks and no end of contentious arguments to confound and distress you. No wonder so many are put off.

MVVM in a Nutshell

In fact MVVM is fundamentally a simple idea. It starts with the realization that your application is an intractable mess in part because you dumped all of your UI logic into the code-behind … with the help of Microsoft tools. You discover the wisdom of pulling much of that logic out into a companion class (and some helpers) leaving the view class to concentrate on looking pretty and capturing user gestures. The view and its companion must talk; you can structure their conversation according to a variety of familiar patterns. Model-View-ViewModel is one such pattern.

The predominance of automated binding is the hallmark of MVVM. You bind a View’s textbox to a string-providing property of the ViewModel so that a user’s changes in the textbox flow to the property and so that programmatic changes to the ViewModel property flow back to the textbox. You bind View events (e.g., button click) to event handlers in the ViewModel enabling user gestures to be conveyed to the ViewModel.

You could write your own bindings by hand, pushing and pulling data and events with custom code. That’s a lot of work. The XAML-based client technologies (WPF and Silverlight) provide a binding framework that requires little to no custom coding. In most cases you can “connect the dots” declaratively and it just works.

Now the View concentrates on appearance and the ViewModel feeds the view. This separation is both clarifying and liberating. You can start looking at that ViewModel to figure out what it should do and how it should do it without also worrying about the visuals. The subsequent examination of ViewModel’s more limited responsibilities sets you on a more sustainable course. At least this has been my experience … and that of the hundreds (thousands?) of other developers who’ve adopted this pattern.

Now you’ll note that I said it “sets you on course”; I did not say it delivers you to the promised land. There are details to work out. If the View looks to the ViewModel for data, where does the ViewModel get that data? Are we talking about data values alone or the business rules that surround that data, rules about data validity, integrity, security, and persistence?

Many of us draw a circle around these concerns and call that circle “the Model”, the third element in the MVVM triad. Our guiding intuition is this: “the persistent data are definitely in the model and if the rules about data are true for this and every possible UI, the rules belong in the model as well.”

If the last name is required, it is required everywhere; that’s a model rule. If gold customers can select from several thank-you gifts, that sounds like UI logic to me … and belongs in the ViewModel. The specifics have to be worked out.

Back To Josh

At this point the discussion risks slipping into theory and theater. We could run to the white board, scribble madly, and count angels on pins without writing a line of code. I love doing that.

Josh has taken another direction which is likely to be both more popular and more effective. He briefly presents his take on MVVM in clear, accessible prose … and then makes it tangible with an extended exploration of a meaty sample application. I’ve read plenty of pattern papers; I don’t remember any that were so thorough … and entertaining … in their examples.

For style and substance this is a remarkable achievement. I truly hope that Josh’s book is widely read because I think it has a better chance of helping WPF and Silverlight developers feel comfortable about MVVM than anything else I’ve seen.

I’m a little worried that some will fear the word “Advanced” in the title. The term “MVVM” is forbidding enough and many folks are still looking for the introduction. I think this book is the introduction they really need.

In what sense is it “advanced”? It’s far from the last word on the subject. It won’t settle any arguments (least of all mine). But it does rise above the introductory level in two important respects.

First, it considers the implications of multiple, cooperating MVVM instances. Most MVVM articles present a single MVVM-triad: a single View and its companion ViewModel and Model.

No application needs a single MVVM triad. If you have a one-screen application, don’t use MVVM; it’s overkill. MVVM is for multi-screen applications consisting of multiple MVVM triads which spring to life somehow and somehow interoperate.

Second, as I noted earlier, Josh doesn’t just talk about the pattern. Most of the pages describe the accompanying game, BubbleBurst, which is implemented in MVVM style. Instead of paeans to principles and their wondrous benefits, you see MVVM in action, warts and all.

BubbleBurst is “advanced” enough to both introduce MVVM and reveal important implementation challenges of multi-MVVM applications. Yet it’s simple enough to be easily understood in the pages of a book or in a few hours spent trolling the code. It’s also fun to play.

Proceed with Caution

I’m about to begin my critical remarks. Before I do, I want to reemphasize my enthusiasm for what Josh has accomplished. This book is a wonderful place to begin your MVVM journey and I found it to be an enjoyable and insightful read. So, hell-yes, I’m recommending it.

On the other hand, it would be a shame if developers stopped their education here. Important elements of the story are missing and many recommendations are seriously misguided in my view. Readers should take Josh seriously but, once they are comfortable with the pattern, they should become aware of starkly contrasting perspectives as well.

Critical Omissions

Four vital topics are neither covered nor illustrated in the code:

• Testing
• Model (the first ‘M’ in MVVM)
• Dependency Injection
• Event Aggregation

Testing

Josh asks “What is the point of having ViewModels?” First among his reasons: “the ability to easily write unit and integration tests for the functionality of the user interface without having to get into the messy world of writing tests for live UIs.”

That’s the last we hear of testing. There is not a single test in the code base. If testing is as important as he says, I’d expect a thorough treatment starting with building a test regimen and showing, by example, how to write good tests. Instead, we get lip service.

I don’t think testability is the only value of MVVM. I couldn’t recommend it for this reason alone. MVVM is a significant architectural investment that introduces complexity. I wouldn’t incur that complexity cost merely to make my UI testable. I’d look for a less onerous route. Fortunately, I think it has other benefits such as clean separation of concerns, improved readability, and easier debugging (try break-pointing XAML!).

But testability is both a motivation for and a by-product of MVVM style and it should have been covered thoroughly in an “advanced” treatise on MVVM.

Moreover, the design would have benefitted greatly from attempts to test it. Many lines of view code-behind are both untestable and harder to follow than they should be. I see early signs of the spaghetti code that afflicts more complex and mature applications … a drum I’ll beat furiously later in this review.

At almost 300 lines, the BubbleMatrixViewModel begs to be tested both for quality and to reveal the expectations latent in its 20+ public and internal members.

The Missing Model

The pattern is called Model-View-ViewModel. Josh’s Model is nowhere to be found. Model and ViewModel are combined and it’s not clear to me how one would disambiguate them.

The absence of Model isn’t blatantly obvious in a simple game like BubbleBurst. The game doesn’t to retrieve or save objects to a persistent store. We aren’t expecting much in the way of business model logic either; there’s nothing of interest to validate or secure. So the lack of a model is not a deep fault of this particular application. Indeed, it’s a convenient way to concentrate our attention on the central issues for most students of MVVM:

• What code belongs in View and what code in ViewModel?
• How do View and ViewModel interact?
• How do you instantiate Views and ViewModels?

Nonetheless, we can’t pretend to provide a comprehensive guide to MVVP without talking about the Model and how it interacts with ViewModel and (to a lesser extent) View. A business application will demand a clear delineation of responsibilities and equally clear communication paths.

I’m not letting BubbleBurst off the hook either. I felt on several occasions that some ViewModels were heavier and more complicated than they should be because an inchoate game model was not independently defined.

Dependency Injection

I appreciate that Dependency Injection frightens many developers. You can discover MVVM without learning about Dependency Injection. Josh can kind of get away without it as long as he doesn’t test his ViewModels and doesn’t have to think about the services required to manage a persistent model.

But I feel you can’t call this “Advanced MVVM” … or even “Intermediate MVVM” … without introducing DI at some point. In Josh’s application, every View and every ViewModel class is known to and instantiated by some other View or ViewModel class. That won’t fly in real-world applications which have Views and ViewModels that are ignorant of each other.

The constructor of the BubbleMatrixViewModel instantiates three concrete bubble components: BubblesTaskManager, BubbleFactory, and BubbleGroup (twice) before making heavy use of them. That’s a lot of dependencies that would make this ViewModel brittle and especially difficult to test.

Taking hard dependencies means Josh can freely assume that ViewModels will always have parameterless constructors and can always be instantiated in XAML, as he often does here. Someone new to MVVM might think this is the norm. It is not.

Josh himself alludes to DI as one of the avenues opened by the MVVM pattern: “[MVVM] means that you can use frameworks like MEF to dynamically compose your ViewModels, to easily support plug-in architectures …”. That’s the last we hear of MEF or dynamic VM composition.

Again, I think we can ignore DI in an introduction to MVVM; we cannot ignore it in an advanced lesson on MVVM.

Event Aggregation: communicating with the unknown

All of Josh’s MVVM triads are statically determined. Their relationships to each other are fixed and well known.

I was surprised to see a ViewModel for one view hold a reference to a ViewModel of a subordinate view. That’s not how I do it; my general rule is that MVVM triads do not hold references to each other. I concede that the rule makes it difficult to organize cross-triad communication. Josh’s approach is simple and elegant … when available.

In real-world business applications, the MVVM triads often can’t know about each other. They are often defined in separate modules without mutual references. Views appear and disappear dynamically. Direct communication is frequently impossible.

That’s why so many so-called MVVM frameworks include some kind of messaging or Event Aggregation infrastructure. You don’t need them to understand MVVM triads in isolation. But you never see an application with a single MVVM triad and you rarely find an MVVM application with statically determined, session-permanent views. No “advanced” treatment of MVVM can neglect Event Aggregation.

View First or ViewModel First?

Here’s an incendiary topic Josh leaves untouched.

I see two ways to ask this question:

• Do I design the View first or the ViewModel first?
• Do I instantiate the View first or the ViewModel first (or does something else instantiate and wed them)?

The answer to both questions need not be the same. For example, you could decide to design the ViewModel first and let the View instantiate that VM at runtime. This is plausibly the approach advocated in BubbleBurst although Josh never talks about it.

Again, it’s important that readers new to MVVM learn that Josh’s way is not the only way.

View and ViewModel Design

In my experience there is a “dialog” between View and ViewModel design. The VM exists to serve a view even as it strives for independence from any particular concrete view. A VM is useless if there is no view that will work with it; clearly the VM developer must heed the imprecations of the View developer.

On the other hand, in business applications the application’s imperatives – what the view must do to satisfy business requirements – are the province of the programmer and are best articulated through the capabilities of the ViewModel.

Therein lies the necessary tension between View and ViewModel design. As a developer my allegiance is with the ViewModel (“the application should do something worthwhile”) but it would be silly to defend that allegiance at the expense of the View (“a good UX is essential to making an application easy to learn and to use”).

I’d have liked to see this tension at least acknowledged and perhaps explored.

Who begets Whom?

Runtime construction is a question apart from the matter of who drives the design. WPF and Silverlight tooling push you toward “View First” construction in which the View instantiates the ViewModel. You’ll often see the View’s DataContext set by a ViewModel instantiated in the XAML as shown here:

  <UserControl.DataContext>
     <viewModel:BubbleBurstViewModel />
  </UserControl.DataContext>

Contrast this with the ViewModel first, code-based approach (not seen in BubbleBurst) in which the ViewModel is assigned to the View’s DataContext as follows:

 view.DataContext = viewModel;

The View First approach assumes that the ViewModel can be created via a default, parameterless constructor. That’s fine for BubbleBurst but wholly unrealistic in business applications whose ViewModels have dependencies (e.g., on the domain model).

Neither WPF nor Silverlight XAML support the dependency injection mechanisms that many of us prefer. I wish Josh had talked about this and perhaps discussed the interesting attempts to use MEF dependency injection (via properties rather than constructor arguments)

I’m not trying to settle these questions. I am saying these questions could have been called out in an “advanced” book on MVVM patterns.

What Belongs In Code-Behind?

Do we allow or prohibit code-behind? That’s a fist-fight topic. Josh looks for a Goldilocks solution, suggesting the amount of code-behind is just right if confined to “logic scoped to the view”.

Such guidance is far too loose in my opinion. It’s not effective guidance at all.

We choose presentation separation patterns in part because we want to minimize testing of the view. We should test any custom logic we write – no matter how it is scoped. The poverty of our testing tools means that, in practice, we will not test the code in the view. Therefore, any code in the code-behind is suspect.

I am open to some code in the code-behind; the “InitializeComponent” method is inescapable. I’ll accept a minimum of “switch board” code in the code behind; direct wiring to a ViewModel member is ok in small doses.

I draw the line at decision logic. I smell a rat when I see a conditional statement of any kind. That’s where bugs breed. Conditional logic is code we should be testing.

I much prefer this rule: “No conditional logic in the code-behind”.

That’s unambiguous. We don’t have to wonder what “scope of the view” means. If there’s an “if” or a “switch” statement, something is wrong. You can detect a violation with automation or visual inspection. It’s a rule that’s easy to follow and pretty easy to abide by … when you know how.

What happens when you don’t follow this rule? You’re lured into the kind of dodgy code that … well we see in the innocent seeming BubbleMatrixView.xaml.cs.

In fact, far from being “best practices”, I’d say that the View code in BubbleBurst represents some of the worst practices.

Look at the HandleMatrixDimensionsAvailable method. This event handler is wired to a UI element in the BubbleBurstView. Unfortunately, the UI element happens to be in a different view, the BubbleMatrixView! One view’s code-behind is wired to members of another view; that’s the road to perdition. That would never pass my review.

I wanted to clean it up immediately. I tried to find the intention behind this practice. Usually I look to the name of a method to discover its purpose. Unfortunately the handler is named for the moment when it is called (“Matrix dimensions available”) rather than for what it does. That’s conventional naming for event handlers but not terribly informative. Unless a handler is brain-dead-simple, I have it delegate immediately to methods with meaningful names. The handler tells me when something happens; the inner method tells me what will be done.

Had Josh adopted my suggestion, he’d have realized that his handler has multiple responsibilities:

• wires the outer window’s keydown event
• picks up the dimensions of the related view’s grid
• starts the game

That’s too much work for code-behind in my book.

When we strive to apply the “no conditional logic in the code-behind” rule, we discover other design problems.

It appears that Undo logic is distributed between two ViewModels. One of them, the BubbleBurstViewModel, determines if undo is possible while the second, BubbleMatrixViewModel performs the undo. Probe further and we find that even the BubbleBurstViewModel test for undo-ability is delegated back to the BubbleMatrixViewModel. The ping-ponging among Views and ViewModels is confusing; it doesn’t smell right.

All the signs suggest the real action is in BubbleMatrixView and its ViewModel.

BubbleMatrixView.xaml is short and sweet.The code-behind in BubbleMatrixView.xaml.cs is another story. Five Region tags sound the alarm. You shouldn’t need region tags in code-behind. Here they barely disguise the complexity of 145 lines of code-behind. That can’t be right.

MicroControllers: BubblesTaskPresenter Example

I was able to refactor the 53 lines dedicated to animating bubble tasks. Most of them went to a BubblesTaskPresenter class, a component Jeremy Miller calls the “MicroController”.

Every MVVM developer should become familiar with the “MicroController” approach.

A “MicroController” performs a narrow UI task on behalf of a view. We usually write one to encapsulate UI logic that is reused across multiple views. That’s not the case here; the application is too small.

We also write MicroControllers to encapsulate UI-specific functionality in a UI-agnostic wrapper. They make it easy for non-UI components, such as ViewModels, to invoke the UI-specific behavior we find in Views. That’s what we need here.

I noticed that the ViewModel holds a TaskManager which raises a PendingTasksAvailable event when there are bubble display tasks. This sets off a sequence of calls between View and ViewModel which collectively “process tasks” queued up by the ViewModel.

Josh’s original code forwarded the event to the View’s code-behind. That code-behind then extracted the VM’s TaskManager and called into it. 53 lines of code-behind are dedicated to this entire process.

After my revision, the view merely provides the ViewModel with a configured MicroController, here called a “BubblesTaskPresenter”:

 _bubbleMatrix.BubblesTaskPresenter = new BubblesTaskPresenter(_bubbleCanvas);

This is what switchboard code should look like. I’d like to get rid of the instantiation but at least there is a bare minimum of logic and no conditionals.

The ViewModel receives the concrete presenter in the guise of a UI-agnostic interface, IBubblesTaskPresenter and wires itself to the presenter’s events.

    public IBubblesTaskPresenter BubblesTaskPresenter {
      get { return _bubblesTaskPresenter;}
      set {
        _bubblesTaskPresenter = value;
        _bubblesTaskPresenter.Completed += delegate { ProcessTasks(); };
        TaskManager.PendingTasksAvailable += delegate { ProcessTasks(); };
      }
    }

    private void ProcessTasks() {
      var task = TaskManager.GetPendingTask();
      BubblesTaskPresenter.PresentTask(task);
    }

IBubblesTaskPresenter uses simple .NET types and looks like this:

  /// <summary>
  /// Interface for a presenter of <see cref="BubblesTask"/>
  /// </summary>  
  public interface IBubblesTaskPresenter {
     void PresentTask(BubblesTask task);
     event EventHandler Completed;
 }

Where did the 53 lines of code-behind go? Into the BubblesTaskPresenter.

What has been gained? Did I just hide the pea under another untestable cup called a “presenter”? Is this merely an academic exercise?

I don’t think so. True, the presenter is dependent upon WPF and upon BubbleCanvas in particular. But it could be tested with a faked canvas and a faked factory. Testing aside, it’s small and focused; it’s easy to see what it does and how it works. It won’t get lost amidst the other motions of the BubbleMatrixView. We might someday use it in another view that presented BubblesTasks.

The BubbleMatrixViewModel has grown a few lines but remains readable and testable; its imported IBubblesTaskPresenter is easy to fake.

The ViewModel no longer exposes the TaskManager. That’s a good thing because the TaskManager is pretty complex piece of machinery with a rich API. We don’t want the View or BubblesTaskPresenter to know about it.

And the View is 53 lines lighter, reducing the anxiety we might justifiably feel if we handed it over to a non-programming designer for styling.

Homework Assignment: if you’re playing along and are looking toward your own application design, you might want to linger over the BubblesTaskPresenter. It is specific to bubbles right now … which is appropriate in the absence of forces that would drive us to generalization.

But you can encapsulate storyboard animations in much this way to both determine state changes within the ViewModel and externalize the animations in “View-space” where they belong. Josh is spot-on when he insists on this separation. The MicroController “presenter” approach is one way to achieve that separation without polluting the code-behind.

Aside: my first presenter was more abstract and general purpose. My generalization was premature. Generalization made the code more complex than it needed to be in this application. I un-factored it to the current version that knows about BubbleCanvas and BubblesTaskStoryboardFactory.

Views Shouldn’t Talk To Views

After clearing out the 53 lines, I was able to focus on what’s left.

I was deeply troubled by the properties of BubbleMatrixView that are exposed to and consumed by the BubbleBurstView code-behind. That can’t be right. Views shouldn’t need custom properties.

If you have to expose a custom view member, you should do so through a view interface. We don’t want any code – View or ViewModel – depending upon a concrete view class. That’s asking for trouble and flies in the face of the testability which justifies MVVM.

It’s also completely unnecessary in this application. As noted earlier the event loop from inner BubbleMatrixView to outer BubbleBurstView leads back around to BubbleMatrixView’s ViewModel. We can cut all that nonsense out and simply update the BubbleMatrixViewModel directly in the code-behind … without any conditional logic.

I refactored as follows:

• Removed 30 lines from BubbleBurstView code-behind
• Added “StartTheGame” to BubbleMatrixViewModel; it combines the functions of setting the dimensions (now private) and starting a new game.
• Removed 61 lines from BubbleMatrixView code-behind which now calls StartTheGame when the canvas is loaded. Region tags are gone.

The BubbleMatrixView code-behind is now a trim 30 lines with no conditionals, down from 145 (80% reduction). Of the 30, exactly 5 of them do work.

My cleanup involved eliminating several null checks. They never did anything useful. First, the variables could not be null in practice. Second, even if they were (e.g., someone broke the xaml), the application would not have functioned anyway. I don’t think a blank screen is superior to an exception. The null checks were about as helpful as empty catch blocks.

Tidy-up BubbleBurstView Code-Behind

BubbleBurstView’s code-behind had one remaining “if” and some other code that I felt didn’t belong there.

Key monitoring logic contained an “if” statement. At first it seemed like key monitoring was a candidate for another MicroController. But there is only one key to handle, the key for Undo. That wasn’t worth the overhead of another MicroController component. Instead I decided to enrich the BubbleBurstViewModel with an Undo method whose signature is:

  public bool TryUndo(bool shouldTry) // returns True if undid

The code-behind calculates whether the key presses amount to an undo request. The call is:

  e.Handled = _bubbleBurst.TryUndo(isUndoKey);

I don’t like the isUndoKey calculation but I’ll leave it for another day.

My final bug-a-boo is the LoadBubbleViewResources method which instantiates a ResourceDictionary and plugs it into the application-level resources.

I expected this to be loaded by the App.xaml itself but I suspect Josh wants to simulate the notion that independent assemblies have their own resources. While the “Bubble.View” assembly is not independent – the App.xaml has access to it through its host assembly’s references –, I’ll play along.

In any case, the view’s code-behind shouldn’t know the details of ResourceDictionary construction and App.Resources update. Time for a MicroController: “BubbleViewResourceLoader” which we can invoke from within the code-behind’s constructor for now. It still stinks but it stinks less.

The code-behind now stands at 41 lines, down from the original 86; only 9 of them do work.

Picky, Picky

I’m a restless complainer. I’ve got a few more bones to pick having nothing to do with MVVM.

Page numbers! You may have noticed that my citations lacked page numbers. That’s because there aren’t any in the printed document I’m reading.

The abundance of Region tags annoys me. Hey, I use them too, especially in pedagogical code where I want to unveil new ideas didactically in the course of a presentation.

I appreciate that they seem to organize the code into nice buckets: “constructors”, “methods”, “fields”. But lately I’ve come to believe that they’re more of a code smell.

If the class is short – as it should be – they interfere with my ability to read and grasp the class at a glance. If the class is long, they suggest that my class implements too many distinct concerns and I’m struggling – vainly – to keep them organized. These concerns would be better delegated to supporting classes. I suspect that automated testing would make this defect more obvious.

I’m not thrilled with the data triggers. Data triggers are not available to Silverlight programmers probably for good reason. The Visual State Manager is the preferred choice for both WPF and Silverlight.

A particularly egregious data trigger can be found in BubbleBurstView.xaml where the GameOverView is collapsed when the BubbleBurstViewModel’s GameOverViewModel is null. Could there be a more obscure way to signal that a modal dialog should be hidden? I suspect a bound IsGameOver property would be both simpler and more obvious.

Where’s The Code?

Josh’s original source is at AdvancedMvvm.com . I’ve temporarily posted my refactored version on my site. I’m not promising it will stay there- this is Josh’s code and I’ll do whatever he wants me to do with it (including burn it) – but it’s there now and I’ll let you know if I move it.

Wrap Up

I have high hopes for the conversation Josh has started with this book.

I see it as one place to begin your study of MVVM rather than as the definitive work on the subject. Some of the practices I find dubious and there are serious omissions.

But I gave it this much attention because it deserves it. It’s a fine piece of writing, a carefully considered example, and a push in the right direction. Josh speaks to a broad audience with an authentic voice they understand and respect. He has my respect as well.

It’s well worth the $15 bucks on the Kindle.

Enjoy!

p.s.: Cross-posted to my other blog on CodeBetter.

Test Soon Development: a case study

I didn’t sleep well after posting about my IoC-agnostic Prism Bootstrapper. Bad enough that I’d deleted the original Prism bootstrapper tests. I added insult to injury trying to justify it:

I really thought this simple task would take maybe a long day at most. Stripping the tests away from the start was supposed to help me stay in the time box. … I didn’t think I could afford [to maintain the tests]. … I wasn’t trying to deliver a product and there is limited value in tests for a version of the Prism RI that is (a) obsolete and (b) irrelevant to your application.

Except it didn’t stay simple and it didn’t take one day. My pride got the best of me. Ok I’m not proud of the code but I didn’t want to be ashamed of it either. So I kept polishing it. The scope expanded little by little; I added Autofac, Silverlight support, and the Prism RI. Four days flew by.

Now I’m invested. Now I care if someone uses it. Without tests I fear no one will. I can hear the scolding voice of Jeremy Miller: “I won’t even look at code that doesn’t have tests.”

I figured going in that setting up the test environment, converting the tests to run with new bootstrappers, and making it all presentable would take a day. That would have doubled the expected one-day project time. That’s ridiculous. So I deleted the Prism tests instead.

It seems a false economy now that it’s a four day project. There’s only a 25% cost if I actually can revive the tests in one day.  I can amortize it further if people use it, improve it, and extend it to more IoC containers.

One more day. How do I justify that? Perhaps, if I document the process and the outcome, I may discover something concrete to say about the utility (or futility) of testing. There’s a blog post in it at the very least.

...

I was right about the one day (see chronology below) … sort of. I squeezed it into one day only by taking shortcuts:

• I worked exclusively with the Prism Bootstrapper test fixture. I did not restore tests for other Prism components nor for the Prism RI.
• I revived tests for Desktop only; Silverlight testing would have sunk the budget.
• I didn’t write any new tests. This being a refactor, the goal was to make all bootstrappers pass the existing tests … and prepare the ground for future tests.

Was It Worthwhile?

Automated testing, they say, confers specific benefits. Did I see them? Let’s review.

Find bugs: “No”. Ok, it revealed one minor omission, insufficient payoff for a day’s work.

Clarify intent: Yes. The tests give me a better understanding of what the Prism authors expected the bootstrapper to do. For example, they prescribed a particular launch sequence and wrote a test to verify that sequence. I may not know why the sequence is important. But were I tempted to write my own bootstrapper from scratch, I’d know to pay attention to this point.

Improve design: No. It did not stimulate design changes to the Systems Under Test (SUT). I still believe that writing automated tests fosters improves design. But I wasn’t writing tests. I was reviving them and, because my ContainerAdapter is supposed to match the UnityBootstrapper API, existing tests were unlikely to provoke change.

Increase confidence: Yes. While tests don’t prove that code works, they do provide strong evidence that certain things work “as intended”. Any evidence beats no evidence. The relief of seeing tests pass, especially tests someone else wrote, is undeniable. I’m hopeful that they improve your confidence in the product.

Speed testing: Yes. Before, if I made any change to the bootstrapping code, I had to run manual smoke tests for each bootstrapper variant. That entailed changing a compiler directive, recompiling, and running the app for each of five bootstrappers; the cycle took 15 minutes minimum.

The automated tests uncovered one bug. With their help, I fixed the bug and confirmed the fix for all bootstrappers in a couple of minutes. That’s a big improvement, for sure, but one speedier repair hardly makes up for the day I put into this. I’m twenty or so cycles away from break even. If someone finds more bugs, or decides to refactor the code to make it more readable or more capable, the payback will come.

Prepare for future  change: Yes. With a test regime and a body of tests, anyone can quickly verify that changes to the CalBootstrapper, changes to any ContainerAdapter, or the addition of new ContainerAdapters preserve certain constraints on how all of the bootstrappers work. Far from perfect to be sure – there are only ~25 test methods per bootstrapper – but we have a foundation in place.

Conclusion

The IoC-agnostic Prism Bootstrapper is better for having automated tests.

That said, the test development effort will be a total  loss if no one uses it. That’s a loss on top of the four days I’d have blown on it in the first place.

It’s a real risk. I am unlikely to use it myself, odd as that seems; Unity is working fine for my existing clients.

On the other hand, I suspect that the mere existence of tests improves the chances that someone will adopt it.

You’d think the IoC product authors might show some interest; their customers want to use Prism. As they improve their containers or come up with new versions, the test-backed CalBootstrapper should make it easer to demonstrate that their enhancements are Prism-friendly.

Patterns and Practices has a stake in promoting a Prism that is open to any container … and they are welcome to take this code and run with it.

Meanwhile, the payback remains uncertain. We’ll have to wait and see.

---

Chronology of Bootstrapper Test Resuscitation

5 minutes		Restore original Prism bootstrapper tests for the UnityBootstrapper; they pass.
1 hour		Refactor test fixtures so not dependent on Unity or UnityBootstrapper; they pass.
30 minutes		Extract Prism’s ancillary mocks (hand coded mocks) into their own assembly; tests pass.
1 hour		Get the tests passing on the “RefinedUnityBootstrapper”. No bugs found; one test failed on a log message change leading to a change in the test. Established a pattern for automated testing of multiple UnityBootstrappers.
2.5 hours		Tests finally pass on the first version of the IoC-agnostic ContainerAdapter. Almost all of the work involves decoupling the tests and mocks from IUnityContainer. Found one bug in my code! I had omitted a Region Behavior mapping. Now can automate testing of multiple Bootstrappers … as long as they require Unity.
30 minutes		StructureMap, the first non-Unity bootstrapper, is using my ContainerAdapter and passing the tests. There was no issue with the SUT. The trouble again was getting the tests to work.
7 minutes		AutofacContainerAdapter test project created, ran green, and checked in. That 7 minute pace augers well for extending the reach of the tests to forthcoming versions of SM and Autofac and to other IoC products.
30 minutes		Package, deploy, download, re-test both the bootstrappers and the Prism RI
~6 hours		Total

An IoC-Agnostic Prism Bootstrapper

Prism from Microsoft Patterns & Practices leans on an Inversion of Control (IoC) container. While nominally indifferent to your choice of container, you have to work a bit to extricate yourself from its default choice which is Microsoft Unity. More so with some containers than others.

The sweat and tears flow in the “Bootstrapper” class with which you launch your application. Most Prism samples feature a Bootstrapper that inherits from “UnityBootstrapper”. “UnityBootstrapper” busies itself with registering and configuring a dozen or so Prism components in a manner characteristic of Unity and it unapologetically takes a hard dependence on Unity itself. You are welcome to write your own Bootstrapper base class from scratch; of course this entails reading and understanding what “UnityBootstrapper” does so you can reassure yourself that your replacement achieves the same effects. Who has time for that?

Almost a year ago I found myself writing a single application with three different client technologies, each dependent upon its own container. The three diverged in minor details – each was a different incarnation of a Container + ObjectBuilder – but the differences were sufficiently great that I felt compelled to unleash the Prism base bootstrapper class from its Unity chains.

Preoccupied with more important business, I took the most expeditious route. I salvaged as much of Prism’s UnityBootstrapper as I could while replacing its direct dependency on Unity library references.

Prism initialization is not something I want to own. When P&P updates its UnityBootstrapper, I want a quick, no-thought pathway to update my IoC-agnostic version.

The result was Cal.Bootstrapper (“CAL” is short for “Composite Application Library”, Prism’s official name) and some supporting types. Cal.Bootstrapper looks like a cleaned-up UnityBootstrapper. It goes through the same motions of registering types and configuring instances albeit with reference to a denuded IoC, shorn of all but the most minimal functionality. This, dear reader, will be my gift to you.

Recently, a few folks have wondered about replacing Unity with a completely different IoC container. The one I hear most often is StructureMap. Its author, Jeremy Miller, has several times entertained the notion of replacing “UnityBootstrapper” with something that reflects glory on StructureMap. Apparently he has better things to do. I guess I’ll tackle it.

The Goods

I’ve bundled everything described in this post into a 2 meg zip, PrismRI-Ioc.zip, downloadable from my company website. Do with it as you will.

Please note: I have no intention of maintaining this; don’t even ask.

No, This is Not the Right Way

My knowledge of StructureMap is deeply impoverished but, from the little I do know, I’m certain that the “proper” course is to abandon “UnityBootstrapper” altogether. But there’s a reason Jeremy hasn’t tackled this and no one seems to have done so for other IoC products either.

Nicholas Blumhardt, author of Autofac, took a crack at integrating Prism with his IoC back in August, 2009. He wanted to do it “the right way,” completely rethinking how a Prism app is bootstrapped and how application modules are written. He describes it here as a “work-in-progress”. I guess it took a little more work than he thought it would and he must have found something better to do because it doesn’t look like he finished.

I hoped to avoid this “road to good intentions” by doing it the “the easy way” which may be the “wrong way” but it works … which beats something that doesn’t.

I decided to change as little of the Unity-oriented Prism app approach as I could. I want to get something done. I want you to be able to convert your running Prism app with minimal effort.

Accordingly, my “Cal.Bootstrapper” mimics the “UnityBootstrapper” and it wallows in the style and ceremony that other IoC products despise.

Is it all that bad? Why am I changing my IoC container anyway? I ought to have a clear idea about what I want from an alternative to Unity. I’m wasting my time if I’m only replacing the UnityBootstrapper with something “better”. Frankly, I don’t care how Prism configures its minimal requirements. If I want StructureMap (or AutoFaq, Castle Windsor, Ninject, etc), it is because I prefer that IoC for registering and configuring my own application components. I may replace the Prism configuration eventually but right now that is neither my motivation nor my first step.

Therefore, although my Cal.Bootstrapper is ugly as sin, it will serve me for the nonce as I hope it serves you and I shall proceed to describe it without further apology. I will let it configure Prism the ugly way using my IoC container in place of Unity after which I am free in my own code to drive my IoC container as it was meant to be driven.

Battle Plan

A Prism application has a Bootstrapper class to configure the container, create and configure the main UI view called the “shell”, acquire application modules, and launch the application.

You are expected to write a custom Bootstrapper that inherits from a base Bootstrapper class provided by Prism (the aforementioned UnityBootstrapper). This base class handles the above inventory of start-up tasks. Your derived Bootstrapper is supposed to extend the base class by overriding certain abstract and virtual methods.

In other words, the bootstrapper is launch karaoke. You add your voice at predefined moments. It’s great if the music suits you. If you’re singing off key, you’re should stop faking it and write your own song.

Tempted as I was, I chose not to quarrel with this approach. In fact, I took a deep breath and then embraced it … for the reasons discussed above.

I decided to replace the UnityBoostrapper with an alternative base class (Cal.Bootstrapper) that does not know about Unity.

Your derived Bootstrapper will provide an IoC container to this base class, wrapped in a vendor-neutral “ContainerAdapter, and delivered to the base class by overriding its abstract “CreateContainerAdapter” method.

I have a few additional goals:

• “Improve” the UnityBootstrapper so I can figure out what it does.
• Make it easy to update my base class when (if) Microsoft updates the UnityBootstrapper.
• Minimize the effort to get my existing app bootstrapper deriving fro the IoC-agnostic base class.
• Minimize the changes necessary to convert my existing Unity-based Prism app.
• “Prove it” with an app Prism-aficionados know: the Prism Reference Implementation.

The original UnityBootstrapper is code only a mother could love. Long methods; much repetition; mysterious names. I can’t replace it until I understand what it does. My first step is to refactor it for readability. This is somewhat a matter of taste but I hope you’ll agree that the my revision, called “RefinedUnityBootstrapper”, is an improvement.

I mustn’t get carried away if I am to satisfy my second subsidiary goal. Too many improvements would hinder re-synchronization with Microsoft’s changes.

I want to prove that my approach works on an “application” that anyone can examine and that would be familiar to Prism developers. The Prism RI is the obvious candidate; that’s what I tackled.

I haven’t looked at the RI in a long time. The version on my disk dates from February 2009. That’s pre-Silverlight 3 and, although I had patched it for SL 3, it’s still old and funky. The UI doesn’t work the way it should: stock positions don’t change when you buy and sell, the “Watch List” feature is broken, the Silverlight version throws a recoverable exception when you sell more stock than you own.

But I didn’t have time to freshen it up, it was good enough for my purposes, and it’s all I’m going to give you.

Swappable Versions

This is a tutorial. I want you to be able to switch among different container implementations without changing the body of the application. I don’t have the time or the interest to do anything fancy.  In the “real” world, I’d make the change and discard the original.

The version of the RI you run depends entirely upon the executing version of Prism RI Bootstrapper class. The class is divided into two partial class files. One of them is constant. The other holds just the code that varies from container to container. The file with variations is called StockTraderRIBootstrapper.Ioc.Desktop or StockTraderRIBootstrapper.Ioc.Silverlight.

The variations are governed by compiler directives at the top. You uncomment ONE of them, recompile, and let ‘er rip. The example below shows the bootstrapper readied for StructureMap.

//***************************************************************************
// *** PICK YOUR BOOTSTRAPPER AND IOC CONTAINER WITH _ONE_ OF THESE DEFINES

//#define ORIGINAL_UNITYBOOTSTRAPPER
//#define REFINED_UNITYBOOTSTRAPPER
//#define UNITY_CONTAINER_ADAPTER
#define STRUCTUREMAP_CONTAINER_ADAPTER
//#define AUTOFAC_CONTAINER_ADAPTER

//***************************************************************************

Each variation takes only a few lines. I set the title text so you can see which version is running on screen (the original three letter “CFI” company name changes to one of “OUB”, “RUB”, “UCA”, “SMA”, or “ACA” to match your selection). For the Ioc-agnostic variations, there is an override of the method that creates the selected Ioc ContainerAdapter. The two versions pinned to the UnityBootstrapper (remember that I “refined” the UnityBootstrapper first simply to understand it better) have a shim method instead.

Incredibly stupid, I know. I’d never ship anything like this. I live with it because it’s a throw away demo and doesn’t merit the effort or complexity of proper decoupling. Or at least it doesn’t merit that attention yet.

The rest of the Prism RI stays constant. It took some work to keep it that way. Not a lot, fortunately; the Patterns and Practices people did a pretty good job of keeping the RI Ioc-neutral.

I won’t bore you with the details of my journey to this blissful place. They are chronicled in “PrismRIChanges.txt”, located in the PrismRI folder next to the Visual Studio Solution file, “StockTraderRI-IoC.sln”. The gist of it is

• Add the Ioc assemblies in a common directory structure
• Add the PrismBootstrapper solution with its Ioc-agnostic Cal.Bootstrapper and the Ioc-specific ContainerAdapters.
• Revise the start-up project, StockTraderRI, to access the above
• Replace references to “IUnityContainer” with “IContainerAdapter”
• Remove module references to Unity.

You’ll do pretty much the same if you convert your existing Prism application to another Ioc container.

ContainerAdapters

The fun is in the ContainerAdapters which implement the following interface:

namespace Cal.Bootstrapper {
  public interface IContainerAdapter {

    IContainerAdapter RegisterInstance<T>(T instance);

    IContainerAdapter RegisterType<TFrom, TTo>(
        ContainerRegistrationScope scope) where TTo : TFrom;

    IContainerAdapter RegisterType(
        Type fromType, Type toType, ContainerRegistrationScope scope);

    IContainerAdapter RegisterTypeIfMissing<TFrom, TTo>(
        ContainerRegistrationScope scope) where TTo : TFrom;

    IContainerAdapter RegisterTypeIfMissing(
        Type fromType, Type toType, ContainerRegistrationScope scope);

    T Resolve<T>() where T : class;

    object Resolve(Type type);

    T TryResolve<T>() where T : class ;

    object TryResolve(Type type);
  }

  public enum ContainerRegistrationScope {
    Singleton,
    Instance,
  }
}

Not so bad if you know your container. Clearly IContainerAdapter must be so dumb that any IoC Container vendor can support it. It needs to be just smart enough for the base Bootstrapper class to configure Prism. Fortunately, this challenge is easily met.

The code tends to be boilerplate. Autofac took a little more work (see below) but still weighs in at only 180 lines.

You also have to write an adaptor that implements Microsoft.Practices.ServiceLocation.IServiceLocator; virtually everyone has done this already.

IoC Containers Demo’d

I’ve tried this approach with three IoC Containers: Unity (Unity v.1.2.0.0, ObjectBuilder2 v.2.2.2.0), StructureMap (v.2.5.4.0), and Autofac (Desktop: v.1.4.5.676, Silverlight: 1.4.4.572).

You should know that I am not an expert in any of these containers. In fact, I just winged it, especially with StructureMap and Autofaq about which I know squat. I figured I could download them, skim whatever documentation they offered, and get something to work adequately for Shmoes like me.

It’s not quite that easy and I have low confidence in my use of these containers. Their authors are invited to lend a hand improving matters … after they’re done laughing.

I wanted to show at least one non-Unity container for both desktop and Silverlight applications. StructureMap doesn’t have a Silverlight version yet (come on, Jeremy). I tried Autofac because it does have a Silverlight version.

I’m glad I did. The StructureMap implementation was a breeze. Autofac was much harder. I’m not trashing Autofac. It just happened to be harder for me, partly due to my ignorance and mostly because of the position Autofac takes on resolving unregistered types.

Major Container “Gotcha”

Container differences matter! One pervasive Prism practice in particular relies on a Unity default behavior that other containers do not observe. It’s serious enough to influence your choice of containers or make your conversion substantially more difficult.

Prism instantiates a lot of its own components with unregistered public concrete classes. For example, somewhere in the bowels of RegionManager initialization, it resolves “DelayedRegionCreationBehavior”. This class is never registered with the container.

Unity is ok with this. When asked to resolve an unregistered public concrete type, it treats the type as if it had been registered and scoped as an instance (i.e., you get a fresh instance each time). Both TryResolve and Resolve behave this same way.

Prism components count on this default Unity container behavior. What if you change containers?

StructureMap has the same default behavior when resolving an unregistered public concrete type. However, it’s TryResolve equivalent, “TryGetInstance”, return null. As I write, I have not detected a consequence of this difference; but I know it’s waiting in the grass somewhere.

You could argue all day about whether a container should return null or return an instance the way Unity does. You could argue whether it should default to Singleton or Instance scope.

Autofac punts. If you attempt to Resolve an unregistered type, it throws an Autofac.ComponentNotRegisteredException with a clear message. “TryResolve” returns null, which is the sensible thing to do.

Aside: the three ContainerAdapter files in my sample implement a “ContainerPlay” class that reveals these differences.

I can’t quarrel with Autofac’s choice but I sure paid the price for purity. I had to hunt down all resolutions of concrete types and manually register them during AutofacContainerAdapter initialization. Unfortunately even that is insufficient. Many class constructors take parameters defined by concrete types. The container will attempt to resolve these and will throw at runtime. There aren’t integration tests to detect these automatically so I have to wander through the app under the debugger. I still haven’t got it to work.

It was easier to get the Prism RI working with StructureMap. I have assumed that the only critical difference is with explicit “TryResolve” calls. I haven’t encountered any optional constructor parameters so I expect StructureMap and Unity to resolve these in the same way.

I searched the Prism and RI source code for “TryResolve” calls involving concrete type. There weren’t many of them. I only found one of potential consequence. The base bootstrapper calls “TryResolve<RegionAdapterMappings>”. Fortunately, it also registers RegionAdapterMappings explicitly so resolution is always successful. Perhaps I was luck. Remember to be careful how you use TryResolve in your application.

The net of it is that containers differ. Prism should not have relied on a Unity-specific type resolution feature. If you have done so in your Unity-based application (as I have), you too are vulnerable.

Where Did The Tests Go?

I removed the test projects from the Prism RI and the Composite Application Library projects that go with it. Why did I do that? laziness?

I really thought this simple task would take maybe a long day at most. Stripping the tests away from the start was supposed to help me stay in the time box. It was going to be non-trivial to setup Prism’s test environment, convert the tests, and repackage for distribution to you. I didn’t think I could afford it.

The point of the exercise was to demonstrate how to liberate your application to use an alternative container. I wasn’t trying to deliver a product and there is limited value in tests for a version of the Prism RI that is (a) obsolete and (b) irrelevant to your application.

Unfortunately, this “simple” task kept expanding in scope; it took four long days. Geez like that never happens.

Now that I’m “done” I wish I had those tests back, the UnityBootstrapperFixture in particular. Debugger testing is painful as we all know. Had I committed to automated tests, I would have written something better than the idiotic compiler-directive-based approach to swapping containers. Sure it would have cost me something – a day? – to get it right, a day I didn’t want to spend. But if someone wants to take this further, the tests would accelerate his progress greatly and imbued his efforts with confidence. Mea culpa.

I wonder if some kind soul will restore them?

Happy New Year, everyone!

PDC 2009 Session Links

I extracted the session titles and links into this PDF to make it easier for me (and for you) to find and navigate to the PDC 2009 session materials. Hope it works for you.

Easing into Prism

I received a familiar email today:

Is it possible to start slowly with this Prism stuff and work my way into it gradually. Can I get something running quickly with DevForce as I know it now? My client has seen a prototype and I need to construct it soon. Then can I come back later and modularize the whole thing behind the scenes?

Here’s my reply:

You do not need Prism to get started!  You can introduce it into your application as you see opportunities that make sense to you. Prism is not a monolithic framework. It is a small collection of free-standing components that also work well together. You can adopt it a piece at a time; some of it you may comfortably ignore.

It’s definitely worth reading about. It’s worth looking at the Reference Implementation and the Quick Starts. Get the Composite Application Guidance book from Patterns and Practices, look at what folks are doing with it and saying about it.

Don’t get hung up on Prism itself. It is guidance first and implementation second. Poke around. See how other people tackle some of the same concerns. You might look at Caliburn, for example.

In other words, take your time and educate yourself.  You don’t have to dive into the deep end.

However, I feel strongly that you should start wrapping your head around Dependency Injection (DI) and Inversion of Control (IoC) containers as soon as you can.  They will completely transform the way you write your application. They will position you for future modularity and make your code ever so much easier to maintain.

Learning IoC is also a great career move.

When you see yourself writing “new XXX(…)” you will learn to stop and ask yourself, “Do I really want to lock myself into the XXX type or would I be better off if XXX were delivered to me instead?”  In recent years my answer has increasingly been to avoid “new” and let the IoC construct and inject the instance I need.

Don’t get carried away now. It’s a shiny new toy. Take it easy. But get used to it.

I know you asked about Prism and here I am waxing on about IoC. I only go on about this because your decision to use or not use DI/IoC is fundamental. Prism (and its ilk) will come along naturally if you use DI; they won’t seem so natural if your application is not DI-ready.

I promise the effort of learning DI (and it does takes some effort) will pay off quickly … and for the rest of your career. The principles are fundamental to emerging Microsoft technologies (see the Managed Extensibility Framework – MEF). Dependency Injection is the primary vehicle for implementation of most modern application development patterns.

You have plenty of IoC containers to consider. Microsoft Unity is one choice but by no means the only one: AutoFaq, Castle Windsor, Ninject, and StructureMap are popular (apologies to any I omitted).

Do not be overwhelmed! These alternatives are vying for your affections with tons of features and you may easily become convinced that DI is super-complicated.

You can ease into DI. Take a look at our “Prism Explorer” (which uses Unity); you’ll hardly notice IoC but it’s there. I limited myself to the simplest Unity techniques. If I remember correctly, I only use constructor injection, I register very few interface types, and I don’t use any of the attribute markup. You can go amazingly far without any advanced IoC features.

In sum, get going on your application and do not let these technologies distract you. They are important and valuable but you don’t want to destroy your business relationship by drowning in technologies that you know nothing about.

Get acquainted with Dependency Injection first and start using it experimentally. Learn about Prism in all that spare time.

Try partnering with someone who has been down this road before. If that is not possible, find a local users group and latch on to other like-minded developers. It’s much more fun when you play with others.

Young @ Heart: The (2007) Movie

Stumbled into “Young @ Heart”, a documentary about a New England choral group performing punk, rock, and R&B covers; the hook is that the members are well over 70 (high of 92).

Sounds like an opportunity for terminal cuteness … “oh those feisty seniors … isn’t it amazing that they can sort of sing at all at their age? … it’s so inspiring!”.  That’s exactly the sense you get from the trailer which is delivered with the typical phone-it-in Fox Searchlight style. Exactly the kind of movie I avoid.

But this one caught me completely by surprise and I’m still roiling with emotions a day later. I’m trying to figure out why.

It should have been utter treacle like one of those sentimental movies about animals, or kids, or illness, or the mentally infirm. Instead it’s a rich character study in which each member of the cast takes you somewhere you’re likely to go … somewhere you want to go … if you’re lucky enough to live that well.

Singing is the glue and reason enough to watch. The songs - well-known pop tunes of youthful love, rage, despair, and hope – confess new moods and meanings as sung by performers far from adolescence.  The juxtapositions can be simple fun as in the raucous “I Want to be Sedated”, given literal treatment by a cast for whom the wheelchair is a serious threat. Or it can rip you open as when Bob Knittle’s rich baritone pleads “I will try to fix you” to the barely perceptible beat of his oxygen machine, counting down toward the inevitable. It would be contrived if it didn’t actually happen, inadvertently, before a packed house of loving fans.

The heart of the movie is the interviews and the lives of the cast as they orbit the rehearsal hall. The authenticity is striking. At a jail house concert the camera lingers on the prisoners’ faces each surrendering a thought of present loss or worrisome future. They are not being entertained (although they think they are); they are transported to the company of someone they miss, of someone they may become.

Death is a looming presence. Close as he is, the cast are no more ready for him than we are.

It’s not great filmmaking. It’s great human material triumphing over the pedestrian. It’s people being themselves, striving for meaning and community, just a few feet ahead of us down the road we travel together.

Patterns and Frameworks

I just returned from the Microsoft Patterns and Practices Summit 2009 which may be my favorite Microsoft conference. Great for exposure to customers with real applications and to the wonderful, dedicated Microsoft developers who aspire to serve them. More on this in another post. I have another point to make here, which is …

Patterns are not Frameworks. We should draw a bright line between Patterns and their Implementations. A pattern only becomes a pattern if and when there are multiple implementations. You don't know a pattern well until you've seen more than one implementation. If there is a single, canonical implementation, the thing is not a pattern, it’s a framework.

“Duh” you say. We know this. And yet it eludes us in practice.

I had numerous conversations about this with folks who work in PnP. The conversation goes something like this:

The Conversation

“We are not a product group. We don’t want to be a product group.

“We’re trying to provide guidance to consumers of Microsoft development products. We are reluctant even to say ‘best practices’ anymore. It’s just the best guidance we can give today, built with time, and commitment, and our access to experts inside and outside Microsoft.

“We’re answering the customer call for Microsoft to demonstrate how to use its own products in real applications.

“This bleeds inevitably into more general questions about how to build applications and soon (and properly) we investigate and describe general design and implementation patterns that go beyond the particulars of Microsoft products.

“It’s not enough to talk about the patterns. We need to demonstrate those patterns … express them in real code that works with Microsoft technology.

“That is when it starts to get away from us. Our guidance, as it becomes code, starts to cover more and more customer cases. It begins to harden and the pattern becomes harder to see behind the bells and whistles. Customers say ‘wow, can I use that in my application?’ and we say ‘sure’. They say ‘it would really help our particular situation if it also did this’ and we say ‘sure’ again.

“Then they want us to give it a name, package it up, support it, develop a roadmap. They want us to productize it … and we start to do that because serving customers is also our mission. And, in our (not so) secret hearts, we all want to ship products … we’re wired that way.

“Then we start worrying about backward compatibility … because customers get angry at us if we change anything

“Then others in the community, especially the Open Source community, get on our case for having sucked the oxygen away from alternative implementations that may be better in many scenarios. They say we have imposed, de facto, a particular implementation of the pattern; people can’t tell the difference between the pattern and what Microsoft did to it. And I see their point.

“And now we have old code (e.g., CAB) that no longer represents our current thinking … that, in fact, embodies practices we would counsel against today … and we cannot walk away from it.

--

I don’t know that there is an easy answer to this one. I’m not willing to place all the blame on PnP either. The customer should know better.

“Best Practices” do not sit still. They evolve, leaving behind a trail of “formerly best practices” to which millions of dollars and entire careers are firmly attached. PnP is pulled in one direction by its responsibility to offer the best current guidance … and by its responsibility to customers who have placed big bets on its past guidance.

Perhaps we need some different rules of engagement.

Explicit Shelf Life

What if PnP put a “Sell By” date on their code? What if it had a shelf life and warning label: “Use with caution. Do not expect support beyond 3 years”. Stamp this at the top of every piece of code and every web page.

Then mean it. If you need to break, you break. Customers need to know that you are serious about your commitment to giving them the best advice you can and you do them no favors by clinging to outdated code.

This happens in OSS all of the time. Dead bodies are strewn everywhere.

You can do this. Microsoft can’t but you can. I think you can wean your customers away from thinking of you as a product group. If that means fewer customers served … that may be for the best … for everyone: your customers, you, and Microsoft itself.

Play Well With Others

My thesis again: a pattern is independent of any one implementation. You don't know a pattern well until you've seen at least two implementations.

What if PnP routinely called out the other implementations of the patterns they describe? Front and center, not as an afterthought. Openly encourage customers to look at these alternatives … if only to gain a contrasting perspective on the PnP guidance.

One of the best ways to learn about the Composite Application Block (CAB), for example, was Jeremy Miller’s “Build your own CAB” series. I assure you I had no intention of following Jeremy’s advice and building my own CAB. It didn’t even occur to me. But that series was the best (unintentional) documentation of CAB I ever found, It introduced me to different ways of thinking about each issue (e.g., Event Aggregation and Dependency Injection). I learned indirectly what CAB was doing and the consequences of my choice. I’m not saying the consequences were all bad either; but there were consequences and I only recognized them in the oblique light cast by Jeremy’s series.

PnP might do more to highlight these alternatives and to encourage discussion about the contrasting choices. People who come to Prism should learn … right there on the home page … that a Caliburn exists. When they come to the Unity home page, they could learn that there are 5 to 10 other popular IoC containers to choose from and find links to learning more about IoC practices.

[I was just on the Unity page. It’s a fine product home page; it is not fine for guidance about Dependency Injection and IoC in general.]

I don’t think that PnP should endorse Caliburn or anything else for that matter. I’m not suggesting that PnP’s output is inferior or that PnP should express less than abundant enthusiasm for its own work.

I am saying that PnP is presenting itself as a product purveyor in a manner that conflicts with its role as a trusted source of guidance. This ensures that it will be trapped into long term support commitments, the kinds of commitments that some people there tell me they do not want.

PnP could say in some unambiguous fashion, “my dear customer, we offer a wonderful implementation of this pattern. HOWEVER … you ought to look at some alternatives, not only when first choosing your approach but also throughout your application development; there’s gold in seeing how others approach these same problems.”

True Open Source

What if PnP could accept contributions from the field? I don’t mean through “Contrib”; I mean directly into the guidance itself.

What if PnP developers actually read some of the Open Source code out there?

I’m pretty sure the legal guardians won’t permit it … in which case “Contrib” is the best they can do.

Conclusion

They call it "Microsoft Patterns and Practices", not "Microsoft Tool and Die".

It is time for customers to pull back from unrealistic expectations of Patterns and Practices. They are not your application infrastructure outsourcer. They are a trusted source of guidance. Rejoice if they provide more but do not expect or demand it.

Patterns and Practice doesn’t need to change what they build. I do think PnP can rethink its commitments to prior output. It can do a much better job of promoting the spirit of exploration and directing its community to external resources that offer contrasting implementations and points of view.

And now I descend from this soap box … in search of another.