Matter of degree, Guest. There's always a degree of rounding off, and giving up. Or else science would have to test each result eternally, and in perfectly identical and somehow simultaneous cases, to test the principle of cause and effect, and generalizations about same kinds and properties of things.
   If science implied perfect knowledge of anything, or complete definition, then it would quit changing its story like an earnest liar and/or an enthusiastic businessman. It implies its particular standards of rigour.
   I think empirical observation is also science, without constructing any tests toward any results. Some people choose a different definition.