[morelisp]

A couple thoughts:

a) The history of mathematics is littered with examples of assumed "obvious" results that turned out to have drastic implications. This is especially true of combinatorics, which was blissfully ignorant to the axiom of choice for so long and now must wrestle with it forever. Strict adherence to some formalism may be justifiably understood, and it's a little shocking to find EWD of all people arguing otherwise for such a pragmatic reason as pedagogical clarity.

b) Comparing 0) and 1) as EWD suggests, the first thing I notice is the sudden appearance of the word "finite" which never appears again in the article (nor in negation; also the inelegant but mostly harmless introduction of "nonempty"). The generalization of the principle to support reals costs it its trivial generalization over infinite bags, which is also often valuable (a classic example being Siegel's lemma).