And is reverted when you remove doubles. Which you should do. Often.
should only be applied to smaller portions and after mesh is finalized, your method is also reverted when you remove doubles (which as you stated you will do often) for this he already has the mesh built the way he wants it, he's already got it rigged as well so weighting won't be a problem. So using my method, in face select mode, C + click once + press y once = done for each triangulated face. While yours on the other hand requires you select each edge individually (or use face select and then unselect the cross section) then mark them sharp which leads to adding and applying a modifier.
Again my method only really works for these smaller pieces, I use edgesplit all the time on my higher poly works which are completely smooth shaded but for something this small and simple I wouldn't bother with it.
