That makes sense and it’s true. What we meant is that static classes cannot be instantiated by a user.

Thank you for your input, we’ll clarify that in the text.

Hey Joe, that’s right. We should have made it clear that we’re talking about instance constructors, not the static ones. Static classes can indeed have one static constructor. We’ve made some changes to make it more clear.

To make sure static classes have a single instance, microsoft designed them to support a parameterless static constructor which is also private, therefore, the static classes can not be instantiated by code; they are instantiated automatically by the framework.

To summerize using causality:
The cause of a static class is to have a sinlge instance, therefore, the effect is it can not be instantiated by code; it is instantiated automatically by runtime

In this db-example, a singleton-class would be preferred over a static class.

A big downside of static classes is that you cannot mock them for testing.

Static classes are usually used for defining extensions on other classes and holding static values like string-constants that are not user-oriented (those should be in the resources where it is easier to provide translations).