Yeah, I do not understand where the author is coming from. One could easily write a helper function in JavaScript that does the same thing that the author's Clojure "cond" does. There is no need to use an IIFE if one does not want to.

Hmmmm... I wonder how you would avoid the expressions from being evaluated without wrapping then in a function (sorry if it's a lame question, I haven't done any JS in years).

Not that there is anything wrong with lambdas, but it's quite a los of boilerplate for something as simole as a cond.

Bingo, IIFEs are and affront to god and nature. Ones in JSX arguably even more so. However, it's really the only way you can colocate all your display code to one part of a function.