Big mistake there, crashfrog. Comparison is a very difficult problem.
Isn't this part of measurement error? If we include measurement error, then it seems to me this concern goes away. Exact comparison is a very difficult problem. But comparison within some definable error is tractable, it seems.
I agree that measurement is not a deductive process; it is a physical one. You could use all the deduction in the world, but without gathering a data point, you're not going to have anything to deduce. Seems a trivial point.
Anyway, how does this tie into the topic? I didn't have the luxury of paging through everything. (But I'm back baby!
)