Problem
Here’s what MSDN has to say about using static classes when it comes to when to use them:
That example, in my opinion, does not seem to cover all of the conceivable uses for static classes. I’ve used static classes in the past for stateless suites of functions, but that’s about it. So, when should a class be declared static (and when should it not)?
Asked by pbh101
Solution #1
In a previous Stack Overflow answer, I discussed static classes: Is a single method class the best approach?
Utility classes with a lot of static methods used to be my favorite. They consolidated a lot of helper methods that would otherwise be scattered about, generating redundancy and maintenance headaches. They’re really simple to utilize; there’s no need to instantiate them or dispose of them; simply fire’n’forget. I suppose this was my first unintentional attempt at building a service-oriented architecture – a slew of stateless services that only did what they were supposed to do. Dragons, on the other hand, will appear when a system matures.
Polymorphism
We can’t, hod. Sure, we can construct a new class and call the old one inside of it if we merely need to add functionality before or after the old method – but that’s just ugly.
Interface woes
For logical reasons, static methods cannot be defined through interfaces. Static classes are worthless when we need to pass them around via their interface because we can’t override static methods. As a result, we are unable to use static classes in a strategy design. By passing delegates instead of interfaces, we might be able to fix several concerns.
Testing
This goes hand in hand with the aforementioned interface issues. We’ll have a hard time replacing production code with test code because our flexibility to swap implementations is constrained. We can wrap them up again, but doing so will necessitate changing a big portion of our code merely to accept wrappers rather than the actual objects.
Fosters blobs
We’ll quickly end up with a massive class filled with non-coherent functionality since static methods are typically utilized as utility methods, and utility methods have multiple functions. Ideally, each class should have a single role inside the system. I’d rather have five times as many classes as long as their goals are clear.
Parameter creep
For starters, that sweet and innocent static procedure might only take one parameter. A couple of new parameters are added as functionality improves. We construct overloads of the method as additional optional parameters are provided (or just add default values, in languages that support them). We’ll have a method with ten parameters in no time. Only the first three arguments are needed; the remaining parameters are optional. However, if parameter 6 is supplied, then 7-9 must also be filled in… We could handle this by taking the needed parameters in the constructor and allowing the user to set optional values through properties if we developed a class with the sole goal of accomplishing what this static function did.
Customers are being forced to build instances of classes for no reason.
le having no further need for the instance? In most languages, creating a class instance is a very cheap process, therefore speed is not a concern. Adding an extra line of code to the consumer is a small price to pay for setting the groundwork for a future solution that is far more maintainable. Finally, if you don’t want to generate instances, develop a singleton wrapper for your class that allows for easy reuse – however this does eliminate the requirement that your class be stateless. If it isn’t stateless, you can still construct static wrapper methods that handle everything while still providing all of the long-term benefits. Finally, a class that hides the instantiation might be created.
Only a Sith is capable of dealing in absolutes.
There are, of course, exceptions to my aversion to static methods. System – True utility classes with minimal risk of bloat are ideal candidates for static methods. As an example, convert. If your project is a one-off with no future maintenance requirements, the overall design isn’t as significant – static or non-static architecture doesn’t matter – but development speed does.
Standards, standards, standards!
The use of instance methods does not preclude the use of static methods, and vice versa. As long as there is a rationale for the distinction and it is standardised. There’s nothing more frustrating than looking at a business layer that’s bloated with numerous implementation techniques.
Answered by Mark S. Rasmussen
Solution #2
Consider what information you’re trying to express when selecting whether to create a class static or non-static. This necessitates a more ‘bottom-up’ programming approach, in which you initially concentrate on the data you’re representing. Is the class you’re creating a physical thing, such as a rock or a chair? These objects are physical and have physical features like color and weight, indicating that you might want to create several objects with varied properties. I’m thinking of getting both a black and a red chair. If you ever need two setups at the same time, you’ll want to create it as an object so that each object can be distinct.
Static functions, on the other hand, are better suited to acts that do not belong to a real-world object or an entity that can be easily represented. Remember that Cforerunners, #’s C++ and C, allow you to construct global functions that aren’t included in a class. This makes ‘top-down’ programming easier. When it doesn’t make sense for a ‘object’ to execute the task, static methods can be employed. By forcing you to use classes this just makes it easier to group related functionality which helps you create more maintainable code.
Most classes may be represented by either static or non-static objects, but if you’re unsure, go back to your OOP roots and consider what you’re representing. Is this a physical item (a automobile that can accelerate, slow down, and turn) or something more abstract? (like displaying output).
You’ll never go wrong if you connect with your inner OOP.
Answered by Despertar
Solution #3
Extension methods are only allowed in top-level static classes in C# 3.0.
Answered by Mark Cidade
Solution #4
If you use code analysis tools (such as FxCop), you should designate a method static if it doesn’t access instance data. There is a performance gain, according to the logic. Mark members as static (MSDN: CA1822).
It’s more of a recommendation than a regulation…
Answered by user25306
Solution #5
For factories, I like to use static classes. This is the logging class in one of my projects, for example:
public static class Log
{
private static readonly ILoggerFactory _loggerFactory =
IoC.Resolve<ILoggerFactory>();
public static ILogger For<T>(T instance)
{
return For(typeof(T));
}
public static ILogger For(Type type)
{
return _loggerFactory.GetLoggerFor(type);
}
}
You might have even noticed that IoC is called with a static accessor. Most of the time for me, if you can call static methods on a class, that’s all you can do so I mark the class as static for extra clarity.
Answered by Rob
Post is based on https://stackoverflow.com/questions/241339/when-to-use-static-classes-in-c-sharp