This is not a math question as much as a process question. For the first time in my (very short) career, I find myself doing one of those messy calculations, where each 'line' of the calculation can spread over a page or three. Essentially all of the calculation is trivial if I'm willing to write some reasonable inequalities in places, and all of these choices are obvious as they're being made, but I am having a rough time keeping the assumptions on term sign required for the various inequalities and the calculations themselves even close to organized, and the copying errors are a nightmare.
Does anybody have any good suggestions on how to stay organized for this sort of trivial-in-theory but messy-in-practice calculation? What do you actually do in these situations? This is especially directed at people in areas like statistical physics or mathematical statistics where these sorts of things show up frequently, and there must be some way of dealing with them. Clever renaming of variables, latexing as you go, good use of Maple...?