CompuServe Messages

<<Some of the "old surveys" actually set survey markers within inches of true position, where the best you can get with normal GPS is 10 meters or so, as I understand it.>> What do you mean by this, Al? "within inches of the true position" has very limited meaning to a surveyor. Your coordinates depend on the datum used, as well as the projection. Your position on the topography is related to the geoid, whch in turn is related to any arbitrary datum ( and hence coordinate set) you wish to define. A point can have literally millions of latitudes and longitudes, all perfectly "true" within their own reference systems. All systems are relative in the end, so what we're really talking about here is the accuracy or "closeness to truth" with which we can define the length and orientation of a baseline within a reference datum. This is made up of the *precision* of the measurements plus their *reliability*. The precision has to have a confidence level associated with it ( ie, "within inches" should encompass a precision measure: ie, 1mm at the 95% confidence level). The reliability is quoted in terms of the probability of detecting an error exceeding a particular size. Both precision and reliability of DGPS measurements are far better than anything achievable with conventional equipment. In 1986 I retriangulated part of northern Brazil with the crude Block 1 satellites. We were getting results better than 3 ppm of the baseline length at 95% confidence even then. We also found a 56 METRE bust in the Brazilian PSAD 56 ( offshore) datum!! Rob