Hm, regarding the tail-optimization slides, the example used doesn't look like a tail-recursive call. I just tried compiling the function he provided using gcc's -O1, 2 and 3, and neither produced a tail optimized call. I rewrote it to match what my idea of a tail-recursive call is, and sure enough, under -O2 and -O3, gcc tail-optimized it. It didn't however with no optimization settings or -O1. I rewrote it as follows:

  long
  fact(long x, long acc)
  {
     if (x <= 0) return acc;
     return fact(x-1, acc*x);
  }

(edit: the whole slide deck is summarized well on the last slide: "If you do an optimization, test it on real world data. If it’s not drastically faster but makes the code less readable: undo it.")