ON THE DESCRIPTION OF COORDINATE REFERENCE SYSTEMS

−=− −     However I question whether this seven-parameter method should have been used for an area as small as Sri Lanka. The transformation parameters are too highly correlated. The land surveying analogy would be that triangulation angles should be between 30o and 150o. Sri Lanka subtends an angle of only 4o at the centre of the Earth. A ten-parameter transformation in which the rotation point is moved from the centre of the ellipsoid to near the middle of the area in question would give a more stable solution, and one in which the translations would more closely represent reality. Should any further work be contemplated, I would recommend a change in approach. Furthermore, I suggest that the precision to which the rotation values are cited is disproportionate to the inherent accuracy of the coordinate transformation, which can be no better than the accuracy of the two survey networks. Three or four significant figures would suffice. I recognise that in presenting this coordinate transformation data the authors are documenting work that had been executed by others, and no criticism of their most informative paper is intended.