What we have done, is moved to responsibility of how a GameUnit is created, to another class.

In the case of this example, a case is used to create new objects. But nobody says this has to be. It could very well be that a particular subclass always creates just one type of GameUnit, or that it creates type A with a lot of different parameters than B.

The logic on how to create a GameUnit is splitted from the class responsible for creating soldiers. The rest of our codebase will work with an InfantryFactory, which is locked (closed for modification), but have an actual implementation inside it (open for extension)

eg:
InfrantryFactory a = SomeClass.GetInfantryFactory(Faction.NOD);

Want a new faction? Create a new subclass, but don’t touch anything else in your code (except the line above ofcourse, to assign your subclass :p)

if we added a new type of unit, we had to modify the subclasses of InfrantryFactory (adding a new case: ..) so why is it such a big advantage that we have to modify those and not InfrantryFactory itself? dont we violate the closed for modification/open for extension principle in either way?

hope i could make my problem clear, keep up the great work + thanks in advance!

P.S.
Article is great too.

I’m happy that everything is doing fine now!

I was hoping for another blog post

Thank you