[larholm]

And robustness is often the static plastic bubble that hides in itself, never accomplishing much in the real world.

State is too often the real world.

Mind you, I like robustness. It's like the cinder blocks I use to build my house, each robust and contained in itself. I just can't disregard state when stacking my cinder blocks on top of each other; the flexing and distribution of load is an emergent state that I have to consider when building each cinder block.

[jerf]

I translate your first paragraph to "I have never used Haskell" in my head.

The idea that Haskell is somehow about not doing anything because it can't do state is just silly. It doesn't pass the smell test. People do not write web frameworks or compilers or anything in a language that "can't do anything" because it can't do state.

Haskell is not about "not touching state". That's just objectively wrong. It may not work exactly like you are used to, you may not like the tradeoffs, but Haskell can do things. I know, because I can make it do things.

[nightski]

So to be clear here I assume you are talking about mutable state. Even if that is the case, it is not about avoiding state at all. It is about managing it explicitly.