It's also worth mentioning that, while the assumption of a linear relation between the dipstick scale and the full quantity might be just about OK for a C172, it's dangerous to generalize. Doesn't work in a P2002, for example. But you can get closer to the truth by running the tank very low or dry, then adding known fuel quantities and noting the levels on a ruler until you hit full. A little bit of graph plotting then gets you the curve you need to make the dipstick.
I agree with Step Turn in that I would never trust a dipstick that I hadn't personally made or checked.
By the way, you'd be surprised at the difference you see between two notionally identical tanks (e.g. LH or RH) when you do the plotting exercise for some aircraft.