Re: RFC Proposal - static modifier for classes

On Thu, Nov 23, 2023, at 1:43 AM, Robert Landers wrote: > I'd venture to say that static classes have some unique benefits over > plain functions. One, you can encapsulate state using private static > variables. That simply isn't possible with regular functions (unless > you use static variables in the functions themselves, which isn't > quite the same thing). It's close enough to the same thing that it has the same problems. As far as I am aware, a static variable in a function is stored in global memory as a singleton. It's logically the same as a global variable, but with access restricted to one function. A static property of a class is also stored in global memory. It's logically the same as a global variable, but with a funny name that has colons in it. It may be access restricted depending on the code. (Engine experts, if I am subtly wrong there please do correct me.) So static variables and static properties are both simply "globals with funny extras." And global-maintained state is well known and understood as a bad idea that hinders testing, refactoring, tractability, maintenance, and everything else developers care about. Using global state, with or without funny syntax around it, is the anti-pattern we should not be encouraging. Generally speaking, even my prototypes eschew globals and statics in 95% of cases. --Larry Garfield