> as long as one can characterise all the test gear correctly it's >possible to compensate readings but it seems to be bordering on the >uncertainty principle sometimes! That's where the imaginary moose come into play. Their state is indeterminate so long as they are unobserved, and measuring them affects the outcome.