I'm trying really hard to understand your argument.
So you believe such a UI API should be immutable (whatever that means in this context), and modelled as a sequence of actions? I think that's fine, but probably not practical in reality. I don't see the point of having a debate about the merits of functional programming vs. OOP in the middle of your post. People are going to build tools around the API that will abstract away whatever semantics you feel are best in order to fit their own needs. So the goal is to build an API that's simple at it's core to implement and use.
Inevitably someone actually has to go and code the adapter layer between the protocol receiver and the actual low level rendering layer (be it OpenGL, UIKit, etc.) and unless you build a system that's easy to understand on that end as well, you're gonna end up with an API no one actually likes or wants to use.
I don't think it's unreasonable to expect that the low-level rendering engine (ie. an implementation in UIKit) maintains some level of UI state. That is - after all - where most of the optimizations will come from. And that's how UIKit, OpenGL and most other rendering systems work. Things generally aren't immutable because it's a.) not efficient, and b.) a nightmare to work with at the lowest levels. It seems to me the appropriate abstraction is build the protocol around a UI state tree, because that's what everyone else is already working with. This is incredibly easy to maintain on either end, and higher level libraries can easily remove the pain of UI state management (a la React, Vue, etc.)
So you believe such a UI API should be immutable (whatever that means in this context), and modelled as a sequence of actions? I think that's fine, but probably not practical in reality. I don't see the point of having a debate about the merits of functional programming vs. OOP in the middle of your post. People are going to build tools around the API that will abstract away whatever semantics you feel are best in order to fit their own needs. So the goal is to build an API that's simple at it's core to implement and use.
Inevitably someone actually has to go and code the adapter layer between the protocol receiver and the actual low level rendering layer (be it OpenGL, UIKit, etc.) and unless you build a system that's easy to understand on that end as well, you're gonna end up with an API no one actually likes or wants to use.
I don't think it's unreasonable to expect that the low-level rendering engine (ie. an implementation in UIKit) maintains some level of UI state. That is - after all - where most of the optimizations will come from. And that's how UIKit, OpenGL and most other rendering systems work. Things generally aren't immutable because it's a.) not efficient, and b.) a nightmare to work with at the lowest levels. It seems to me the appropriate abstraction is build the protocol around a UI state tree, because that's what everyone else is already working with. This is incredibly easy to maintain on either end, and higher level libraries can easily remove the pain of UI state management (a la React, Vue, etc.)