C# – Using static classes as namespaces

cnamespaceobject-orientedstatic-typing

I have seen other developers using static classes as namespaces

public static class CategoryA
{
    public class Item1
    {
        public void DoSomething() { }
    }
    public class Item2
    {
        public void DoSomething() { }
    }
}

public static class CategoryB
{
    public class Item3
    {
        public void DoSomething() { }
    }
    public class Item4
    {
        public void DoSomething() { }
    }
}

To instantiate the inner classes, it will look like the following

CategoryA.Item1 item = new CategoryA.Item1();

The rationale is that namespaces can be hidden by using the "using" keyword. But by using the static classes, the outer-layer class names have to be specified, which effectively preserves the namespaces.

Microsoft advises against that in the guidelines. I personally think it impacts the readability. What are your thoughts?

Best Answer

Using static classes as namespaces defies the purpose of having namespaces.

The key difference is here:

If you define CategoryA, CategoryB< as namespaces, and when application uses two namespaces : 

CategoryA::Item1 item = new CategoryA::Item1(); 
CategoryB::Item1 item = new CategoryB::Item1();

Here if the CategoryA or CategoryB is a static class rather than namespace, the usage for the application is almost same as described above.

However, if you define it as a namespace and application uses only 1 namespace (doesn't include CategoryB), in that case application can actually use following 

using namespace CategoryA; 
Item1 item = new Item1();

But had you define CategoryA as a static class the above is undefined! One is forced to write CategoryA.something every time.

Namespaces should be used to avoid naming conflicts whereas class hierarchy should be used when class grouping has some relevance to the system model.

Related Topic