O.k. Math folks like to calculate things as precisely as possible. However, there really is a limit on how precise one can measure something, even with good temperature controls. So, what is the upper limit on the practical number of digits one needs for the approximation of pii?
Take the example where a disc is created whose radius is 10 meters, and is created out of a sheet of graphene. The potential limit of precision, then, becomes bounded by the size of the carbon atom.
Now, in my world, I have had occasion to measure something I've machined whose diameter is 48", and I typically measure to 0.0001" (but not something that big to that level of precision). So, using those extremes, my limit for pi is only seven digits (to avoid any round-off errors).
My question for you is to justify a greater level of precision, and express how many digits that entails. I'm going to go out on a limb, and say that anything beyond 12 decimal points is really a waste of time for practical applications.