[a short digression]

huh. in one of my proofs, i forgot to fix an $\epsilon$;
i guess the referee assumed it was given?

[thinks]

---

that reminds of the time that a ph.d. student acquaintance of mine was giving a seminar talk. before estimating a certain integral, he fixed a positive parameter $\epsilon$. by the time he finished, there was silence for a moment, and her/his advisor asked:

"wait: where did you use $\epsilon$?"

then everyone scanned the lines of estimates, but nobody could find that symbol anywhere.

"well," the student replied, "it's always good to fix an $\epsilon$, you know ..?"

.. to which the crowd laughed. (-: