Oct 01, 2011 21:45
I'm supposed to talk about differentials as a tool for error propagation in class this week. I could use some real examples, because the ones in the textbook aren't fooling anyone.
So, scientist and engineer types: Are there important quantities in your work that have to be measured indirectly? If so, can you tell me anything about them? In particular: What do you measure instead? What's the formula? What are typical values and uncertainties for the measurements? (And, for what it's worth, how do people keep track of error bars in your field?)