Things you should and shouldn’t calibrate!
http://www.youtube.com/watch?v=Mbn1ckR86Z8
Oh thank you for this!
Call me stupid, but I have had not printed a single calibration cube yet. I couldn’t see the point of it. I printed what I wanted and then if that was not the size or precision I needed, I checked the steps and adjusted them as required, which I did once only.
Other than that it is always a case of judging inside and outside diameters against slicing parameters depending on print.
@Lucas_Fowler the one you see in the video is actually the first one i printed, too.
@ThantiK it had to be done.
Nice video! You may want to point out that back in the Darwin and Sells days - and even the early Prusa days, we were printing our gears & pulleys, so the tolerances were only as good as the printed parts. I’m not even going to wander down the road of trying to print smaller than GT5 toothed gears. 
This made calibrating your printer a unique experience that was only loosely based on the maths that defined the system. There were loads of ways of doing that calibration, and most a lot more accurate than using a 10x10 cube!
Today we are spoiled - and lucky - that we can buy re-enforced GT2 timing belts, machined pulleys, and gears that make theoretical accuracy much closer to reality for a price most can afford.
@John_Driggers been there, done that. Don’t want to do it again. That Sells i show at http://www.youtube.com/watch?v=Mbn1ckR86Z8&t=0m34s was actually the first printer i built.
However, shoddy pulleys don’t influence your large-scale steps per millimeter. They might cause some local errors as they wobble about, but other than the driver’s microstepping and the stepper angle, only the number of teeth on the pulley and the belt’s pitch (which was pretty spot-on with steel-wire reinforced T5) influences scaling.
Aye, and we 'ad ter werk down 't mill fer 22 hours a day and pay 't miller fer’t privilege ur doin so!
I remember how miraculously things changed when I ditched my DC extruder motor and converted to a stepper. Before that the dimensions were mostly stable, but the fill changed depending on how long the machine had been running, the temperature in the basement, what other loads were on that leg of the power supply…
@Thomas_Sanladerer
My first set of pulleys were over extruded enough that the effective diameter was not what it should have been - so just a bit of a scaling problem. I’ll have to go see if I can find 'em - I suspect I still have most of the bits! 
@John_Driggers that’s why you never use printed pulleys.
Times change @ThantiK - I hate to think of what I paid for my original GT5 belts.
@John_Driggers , I’ve been doing this since the original cupcake. You never use printed pulleys. That’s pretty much not changed…I cringe about my old T5 belts, but I used proper T5 pulleys.
I’m still using T5 on my old i2 workhorse, tbh that printer is probably as good as it’s going to get. I’m taking the quest for quality with me to ventures new.
I’ve also never printed a calibration cube.