i don't know why but statistically the number of times i've written code to BFS on a binary-tree given the day-hour experience i've had, is zero.So why don't interview patterns change too. If we expect the person to design,code,execute with you , why don't we check those aspects instead of pure-academic questions on integrity of a tree-search for high-performance solutions delivered in a 10-minute conversation. Also…
It irritates me because it's not even remotely "academic". Category theory is academic, this is mundane things. It probably is the bare minimum one has to know if they hope to work on anything interesting.