When using a digital level for bluebook work, one must determine the inclination correction at the beginning of each day. The instrument then applies this to all readings. So, as you ask, what is the point of requiring such tight agreement between foresight/backsight distances? In fact, if the correction is above a certain amount (10"), you must re-do the determination. We were consistently getting a value (-20") that was outside the limit, but it was consistent. So, if it is being applied to the observations, what is the point of having the limit? We did have the instrument adjusted, which brought it down to around 7".
(rhetorical questions)
John,
My guess is that it was a limit that was based on the instrument's original specs as published. As long as it can be corrected, then there's no difference & I agree with you.
Precise leveling - SUMMARY
This thread got some good responses, but the general consensus for using a stiff leg tripod vs. an extension leg tripod seems to be so that you don't have any leg slippage.
On a purely logical basis this just doesn't make much sense to me because I suspect leg slippage would be a lot more of a problem with a robot that is applying all sorts of motion and torque tot he tripod. And yet we use total stations & robots day in and day out on extension leg tripods.
Earl alluded to the fact that a stiff leg tripod in and by itself does not guarantee quality.
I suspect this is an outdated standard (although not a bad one) that dates back to the 1800's and first half of the 1900's where precise levels were quite large and heavy and didn't have automatic compensators. And at that time most extension leg tripods were fairly light weight and designed for smaller instruments. Those extension leg tripods simply couldn't handle the weight of the level.
So, while I didn't get any good solid answers to the original question, I do appreciate the responses and thoughts and will start working on my next puzzler.