I agree wholeheartedly with the main thrust of your argument, but I think it is slightly disingenuous to put confusability of goto and tail calls in the same l

I agree wholeheartedly with the main thrust of your argument, but I think it is slightly disingenuous to put confusability of goto and tail calls in the same league. Yes tail calls are gotos, but it is at an order of magnitude better than a goto at conveying the meaning of the code snippet to a human. This is anecdotal, but some reasons, goto labels are not as well named as function names. With a goto it is more difficult to guess what data the destination code is expected to touch. Tail calls are an improvement on both the counts. I often do not have to break context from reading a piece of code to start reading the piece of code that the program counter jumps to (unless it is a very local jump). The arguments passed by a tail call and the name of the function that has been tail called often makes (i) breaking code reading context less necessary and (ii) make larger semantic jumps remain within realms of easy human reasoning. With gotos, the jumps better be within 20 lines of text.