I’ve always heard it as basic quorum before Raft was a thing. One clock and you never know if it’s right or wrong. Two clocks and you’re worse off than one in case they disagree. Three and at least you can (probabilistically, anyway) reach a rough consensus good to the deviation of the clocks not exceeding some threshold in comparison to the other two.
Why does it matter to reach strict consensus with a quorum of half plus one in the case of a continuous variable such as time of day? Wouldn't averaging work (regardless of the parity of the members)?
I think it's assuming you have no idea of the time and one of the clocks may be very wrong, like several hours off. So you'd have two clocks close to each other and one far off.
Yessir.
Source?
An assertion with so many conditional assumptions that it's pragmatically meaningless.