Incorrect, reread the first two paragraphs:
> Of course one can define functional programming so that no local mutable state and no side effects are possible, and then point out the obvious disadvantages. But that's perhaps a "no true Scotsman" kind of argument. If you would define object-oriented programming with the same strictness, everything that is not an object and uses mutable values would be forbidden. Even something such as y = sin(x), copy-on-write or returning constant objects or containers from a function would be off-limits. That is, I am asking with a very pragmatic definition of FP in mind!
> What I am wondering is rather, do you observe pragmatic functional programming as John Carmack described it here as valuable? And of course this question goes also to people who have tried it to some extend, as a purely theoretical discussion would be boring. I am interested in good examples!
It is only easier to reason about when the data you give the function is limited. You can always pass the world into the function and take it as an output, then it no longers matter. Effect systems are also possible with objects as well, they just aren't a feature of functional languages.
But if you need a kilotonne of developers stick with a mainstream language.