I think this is compelling programming advice, but "cognitive load" isn't adding anything. The problem with popularizing a term like cognitive load is that it just becomes a new more scientific term for an existing concept in folk psychology. It can then be applied, detached from any experimental psych, to anybody's personal bugbear.
"This programming paradigm is bad because cognitive load" becomes identical to "this programming paradigm is bad because it isn't simple" and then simple is such a fuzzy concept that you can define whatever you like as simple and whatever you don't as having high cognitive load.
> then simple is such a fuzzy concept that you can define whatever you like as simple and whatever you don't as having high cognitive load.
Good observation. That is absolutely rampant in online programmer discussions/flame-wars. Which is why Rich Hickey's classic presentation Simple Made Easy, although not the last word on the topic, at least tried to bring some objectivity to what simple is.
Which is why engineers who intend to use the term in their discussions shouldn’t dumb it down or loosely define it in their own words, and cite the definition from authoritative sources, say reference texts.
And whenever it is being used as an argument to reject or approve code, just saying “because cognitive load” should not be accepted as enough. Instead, there should be an accompanying explanation for what exactly in the code raises the cognitive load and what mechanisms come into play that creates the cognitive load according to cognitive science. (Note: I’m using “science” here as opposed to just “psychology”, because the ana/physio of human memory is not exclusive to the domain of psychology.)
"This programming paradigm is bad because cognitive load" becomes identical to "this programming paradigm is bad because it isn't simple" and then simple is such a fuzzy concept that you can define whatever you like as simple and whatever you don't as having high cognitive load.