Thread-Safe Lazy Singleton Implementation in .NET 4

It has been a while since I needed to develop a singleton in .NET, and now that I am working in .NET 4, I thought it was a good time to revisit the implementation and take advantage of new language features.

So without much fanfare, here it is:

 1Public NotInheritable Class MySingleton
 2 
 3#Region " Thread-Safe Lazy Singleton "
 4 
 5    Private Shared _instance As Lazy(Of MySingleton) = New Lazy(Of MySingleton)(Function() New MySingleton())
 6 
 7    ''' <summary>
 8    ''' Hide the constructor
 9    ''' </summary>
10    ''' <remarks></remarks>
11    Private Sub New()
12        ' nothing to do
13    End Sub
14 
15    ''' <summary>
16    ''' Singleton instance
17    ''' </summary>
18    ''' <value></value>
19    ''' <returns></returns>
20    ''' <remarks></remarks>
21    Public Shared ReadOnly Property Instance As MySingleton
22        Get
23            Return _instance.Value
24        End Get
25    End Property
26 
27#End Region
28 
29End Class

The .NET 4 Lazy class by default ensures that the singleton is thread-safe.

If you use the Lazy class and do not pass in any parameters, it will construct the type automatically for you… which means that the class must have a public constructor method with no parameters. By passing a value factory delegate into the constructor of the Lazy class, we can keep our constructor private and thereby force others to use our Instance property and truly make a Singleton.

Guidance for Using Optional and Named Parameters In C# & .NET 4.0

Leading up to the release of Visual Studio 2010 and Microsoft .NET Framework 4.0, there have been a few posts about how C# 4 now has parity with Visual Basic with regards to the ability to use Optional Parameters. Unfortunately, some of the examples that demonstrate this feature are… well… somewhat disturbing to me from a code quality perspective. In addition there is a lot of talk about how ‘cool’ this feature may appear (from 60,000 feet!), but there is little discussion about recommended guidelines for using Optional Parameters and when it is safe to do so.

Let’s start with the code analysis rule CA1026: Default parameters should not be used. Even in Visual Studio 2010, this rule still exists! Previously a large argument against using Optional Parameters was that code in VB.NET could use them but code in C# would not be able to use them. (This was especially annoying with COM interop!) Before C# 4, a C# coder would have to explicitly provide an argument for each Optional Parameter that was defined (annoying). While this is no longer a concern for C# coders, it is however still valid for consideration when creating methods that may be called from other .NET languages that don’t support the usage of Optional Parameters.

Recently this video on Named and Optional Parameters was released. While the author does a reasonable job conveying the essentials, the scenario is simply a code smell. From a clean code perspective, instead of passing a bunch of search criteria parameters to the ‘Search’ method, the SOLID principles should be used and the criteria should extracted into its own class… This would then mitigate the need for optional parameters in the first place.

In ScottGu’s blog post Optional Parameters and Named Arguments in C# 4 (and a cool scenario w/ ASP.NET MVC 2) [respect to ScottGu btw for your many good works :)], the “cool scenario” replaces a Nullable parameter named ‘page’ with an optional parameter that has a default value, citing that the Optional Parameter and default value on this public method essentially makes the behaviour of the method expressed “more concisely and clearly”.

Now, I agree that the code may appear slightly more ‘concise’, but I think it is quite arguable that it is more ‘clear’.

Side note: is ‘page’ a 1-based number or a 0-based index? Isn’t a page in the real world usually a 1-based number? e.g. “Page 1 of the search results”. In the spirit of the example code, perhaps the parameter really should be renamed to “pageIndex” - I think that would make the method more clear…

Getting back on track though: by inserting a default value into the method signature, a whisper of the internals of the business logic of this method is hard-coded into the public API.

Read that again… there is subtlety in there that leaves the feature of Optional Parameters open to unwittingly known abuse by the majority of the population! And it is the lack of clarity due to this subtlety that greatly concerns me from a code quality perspective.

What are the subtle caveats?

  1. As per the CLI specification, the default value must be a compile-time constant value.

  2. When compiled, the default value is emitted into a DefaultParameterValue attribute as metadata on the parameter within the method signature.

  3. When the compiler is compiling the code that calls the method, the compiler reads this value from the DefaultParameterValue attribute and actually hard-codes that value into the IL at the call-site. In other words, every line of code throughout all applications that call that method have this value hard-coded in their calling assembly.

As an example, if you are using say Version 1.0 of the assembly with the method which specifies a default value of 25 for the parameter, your code will be compiled and have the value of 25 hard-coded into your assembly. If you upgrade to say Version 1.1 of the assembly with the method (which now specifies a default value of 26 instead of 25), and if you do not recompile your code, your code when executed will pass the method a value of 25. The public API has been broken, and almost everyone would be unaware! (I found that this information is described in much more detail in C# 4.0 FEATURE FOCUS - PART 1 - OPTIONAL PARAMETERS and due to my concern about the age of the post I actually verified it myself with the help of Reflector…)

Anyone else flashing back to why “public [static] readonly” values should be used instead of “const”? It’s the same argument! The const value is treated as a literal and is hard-coded into each assembly that uses the value. (Conveniently, for more information, you can read THE DIFFERENCE BETWEEN READONLY VARIABLES AND CONSTANTS)

  1. Alternatively, if you as the caller of the method were to actually name the parameter in your calling code, and if you upgraded to the next version of assembly with the method which has had the parameter renamed, your code will not compile! Why? Because when using Named Parameters, the parameter name itself actually needs to be considered part of your API!

What should be the guidelines for using Optional and Named Parameters?

In general, I would suggest that Optional Parameters (just like the usage of the const keyword) should only ever be used with constant values that will never ever change over time - for example, mathematical constants… like the value of PI (although I don’t know why you would make PI an Optional Parameter… but you get the idea nevertheless). Generally and unfortunately however this will typically exclude any ‘constant-like’ values in your business domain. Values in business domains, while seemingly constant for now still can change over time as the business changes!

As always there are going to be a few perspectives on this. I find that it is useful to categorise development into two areas: (a) application development, and (b) reusable library development.

Especially when developing reusable library code, extreme care needs to be taken when creating and maintaining public APIs. I would suggest that Optional Parameters should NOT be used in public APIs of reusable library code… instead revert to the usage of method overloading, or redesigning your API.

However, in application development, where all the code is highly likely to be recompiled (and often), perhaps it isn’t so bad to use the Optional Parameters?… that is, until a team member decides to clean up the code by doing some refactoring and move that method into a reusable library… and then it all begins!

Summary

So for safety, clarity, simplicity and consistency across all code-bases (especially with the widely varying technical capabilities of developers), perhaps the best practice guidance for using Optional and Named Parameters should be:

  • Never use Optional Parameters in public methods; and
  • Only use Optional Parameters with default values that are constant in time (like mathematical constants and not apparent business domain constant values).

Program Specification Through the Mechanism of Automated Tests

My previous findings regarding the testing of public methods versus private methods opened the door to allow me to better understand that the mindset of a developer performing unit “testing” is different from the mindset of a developer “specifying program behaviour through the mechanism of automated test methods”. The difference is subtle but important.

In the concept and terminology of Behaviour-Driven Development, Design, Analysis and Programming, the resulting program code is simply an implementation of previously specified behaviours. Those said behaviours are initially defined at the business level via the process of requirements gathering and specification, and can be expressed in the business’s language as a set of scenarios and stories. These requirements can then be further analysed and broken down into logical and physical program specifications.

Regardless of whether a test-first approach to development is taken, by the end of the construction task, there should be a set of test cases (automatic and/or manual) that were successfully executed and that covers all of the logic specified in the program specification. In other words, by the end of the construction task, the developer has verified that the program behaves as specified.

Typically a program’s specification is written in a document somewhere which is divorced from the location of the actual code - so great discipline and effort is required in order to keep the specification and code in synch over time. Depending on the company and financial constraints of the project, this may not be feasible.

However, what if we could remove this boundary between the document and the code?

Automated test cases serve a number of useful purposes, including:

  • Behaviour verification;
  • Acting as a “safety net” for refactoring code;
  • Regression testing;
  • Professional evidence that one’s own code actually does what one has said it does; and
  • Acting as sample code for other developers so they can see how to use one’s code.

What if the automated test assets were the actual living program specification that is verified through the mechanism of test methods?

This concept has been recognised before, and the libraries JBehave, RSpec, NBehave and NSpec have been the result. While I honour the efforts that have gone into those libraries and the ground-breaking concepts, I do not necessarily like the usage of them.

In fact, all I want right now is to be able to express my specifications in my MSTest classes without the need for a different test runner or the need for another testing framework. In addition, I want other team members to easily grasp the concept and quickly up-skill without too much disruption.

While a DSL might be more ideal under these circumstances, I write my automated test assets in the C# language. Working within that constraint, I want to express the program specification in test code. With special note to the agreeable BDD concept that "Test method names should be sentences", I devised the following structure for expressing program specifications as automated test methods.

 1namespace [MyCompany].[MyAssembly].BehaviourTests.[MySystemUnderSpecification]Spec
 2{
 3    public class [MyPublicMethodUnderSpecification]Spec
 4    {
 5        [TestClass]
 6        public class Scenario[WithDescriptiveName]
 7        {
 8            [TestMethod]
 9            public void When[MoreScenarioDetails]Should[Blah]
10            {
11                // Test code here
12            }
13
14            [TestMethod]
15            public void When[AnotherSubScenario]Should[Blah]
16            {
17                // Test code here
18            }
19        }
20    }
21}

This structure allows for the naming of the public method that is under specification (the parent class name), a general scenario with a description of the context or given setup required (the nested MSTest class name), followed by the test method names which further narrow the scenario and provide descriptive details of the expected behaviour.

The test names look neat in the Test View in Visual Studio (when the Full Classname column is visible), and are readable just like their corresponding stories.

The full class name in this instance would look like:

1[MyCompany].[MyAssembly].BehaviourTests.[MySystemUnderSpecification]Spec.[MyPublicMethodUnderSpecification]Spec+Scenario[WithDescriptiveName]

With automated spec/test assets/classes structured in this fashion, instead of developers performing unit “testing”, developers can instead embrace the mindset of “specifying program behaviour through the mechanism of automated test methods”. In turn, developers should be more focused on their tasks and it should also lead to higher quality code bases.

Oh, and I haven’t even mentioned how much easier it would be for a developer new to a project to learn and understand the code base by looking at the test view and reading all the scenarios that have been specified and coded…

If you wanted an even better way to define your program specifications and test assets, then I highly recommend SpecFlow.