There are other advantages to using static factory methods to wrap object construction instead of bare constructor calls.
As well as allowing for different methods to use in different scenarios, with more relevant names both for the methods and the parameters and without the constructor having to handle different sets of arguments of different types:
* You can do all your input validation before attempting to construct the object.
* The object itself can bypass that input validation when constructing new instances of its own class, since you can ensure that it knows what it's doing.
* With input validation/preprocessing moved to the factory methods, the constructor itself can often be reduced to "set these properties to these arguments", meaning the constructor promotion syntax becomes more useful.
* Having been hidden away from users, the constructor's signature can be a bit uglier without becoming a pain for them. Heh.
* Static methods can be lifted and passed around as first class closures, to be called in the normal fashion wherever functions can be called, without the special "new" syntax.
* The factory method need not return a new instance of that exact class. It could return a pre-existing instance that would do the same job as the new one would (especially useful in the case of immutable "value type" objects by reducing duplication); or a simpler or more specific subclass to do the job with less overhead than a more generic instance of the original class. Returning a subclass means LSP still holds.
