I believe that is a set of geodetic transformation parameters that maintains sub-millimeter precision in the transformation, rather than a sub-millimeter geoid, or sub-millimeter positioning system...
Every millimeter counts, in the Nether Regions...
One of my hobby interests is machining.?ÿ As a rank amateur, I struggle to hold tolerances tighter than about 0.005 inch, but some of the more experienced and talented guys deal in microns.?ÿ It's pretty impressive what a knowledgeable person with the right equipment can accomplish.
?ÿ
Sounds like another great sales pitch from people that don't actually use the technology they didn't develop or design....
?ÿ
Trimble is just announcing that they have incorporated it into a newer release of the Coordinate Systems Database.
I don't think there was any sales pitch involved. My guess is the folks over at the Nederlandse Samenwerking Geodetische Infrastructuur, who developed the transformation, are pretty knowledgeable about geodesy, as well as the capabilities of modern GNSS equipment...
?ÿ
That quote in the OP seems like a perfectly reasonable statement. They're not talking about accuracy of positioning methods (total station, GNSS, whatever...) but rather accuracy of computational algorithms. It looks like RDNAPTRANS is a software library for converting between the Netherlands' "RD" plane coordinate system and ellipsoidal coordinate systems. Some geodetic problems don't have closed-form mathematical solutions (i.e. can't be expressed in an equation) and have to be computed by other methods, such as approximate, iterative algorithms. There can be tradeoffs in accuracy vs complexity of the algorithms and the computing resources needed by the algorithms (less of an issue today vs decades ago, but still not to be ignored). A geodetic software system designed to keep its computational errors significantly smaller than the inherent accuracy of the observations, measurements, or coordinates processed through it is a good thing.
The accuracy of an algorithm can be determined by comparison against some other, rigorous, "expensive" method, or perhaps checked by round-tripping through the algorithm (strictly, a pair of algorithms), i.e. doing a forward calculation and then running that result back through an inverse calculation.
Examples of attention paid to accuracies of geodetic algorithms can be seen in the work of USAF / NGS geodesist Thaddeus Vincenty, who developed algorithms for computing ellipsoidal forward and inverse problems with millimeter-level accuracy for nearly antipodal points on the Earth (~18,000-km geodesics). Further work was done in recent years by Charles Karney of SRI. Other, simpler methods, such as Bowring's, yield similar accuracy only over "short" lines, e.g. up to about 150 km.
For some further info, see: https://pdfs.semanticscholar.org/de80/1fbaf381d0c10186042348ad175573d05537.pdf
