This is a great thread though misguided here and there. The Delury paper cited by Donald Strickert clarifies a great deal of the confusion. I suggest anyone reading this read that.
Stats 101 might teach the issues of reporting versus computing and of real-world relevance, e.g. WalMart revenues down to the dollar, but Engineering 101 would have also taught statisticians that significant digits DO matter and that reporting anything with greater precision than was inherent in the underlying measurements is, at best, misleading. The Delury paper cites one such dilemma.
-------------------------------------------
Thomas Taylor, P.E.
Mathematical Statistician
CDC
-------------------------------------------