One thing that makes programming difficult is how much context needs to be held in the live part of the brain. Past N items (let's say 4), the brain has to swap

One thing that makes programming difficult is how much context needs to be held in the live part of the brain. Past N items (let's say 4), the brain has to swap those values and it makes programming much much slower. Most of the article is spent explaining this sideways; with Haskell you need to hold less aspects in the head because they are either eliminated entirely (purity) or can be deferred to the compiler (types), ... What we learn is that Haskell is easier to implement tree-like structures. I think this is what most language proselytes are trying to convey in their articles. But don't talk about explicitly. Inevitably they will pick a task that is easy to achieve in the language because the developer environment aligns well for that use-case, and then let the reader infer that this applies to all programming tasks. Basically we are trying to benchmark humans, without building a proper model of how humans work. The next evolution in programming will be done by properly understanding how we work and interact with the computer.