-> You may be right; I haven't analyzed this problem in all its details.
-> One could, of course, draw a few sample triangles on graph paper,
-> plug the co-ordinates into the formula I have given, & see whether
-> observed results (measured by protractor) tally with theory.
Or, much simpler, see if the results your method produces are identical
with those that the other two methods (subtracting ATNs, or using the
cosine rule on the triangle) predict. I've already got these two methods
programmed up. They both work, and produce answers that are identical
apart from occasional round-off errors in the Nth decimal place.
I may try it someday...!
                              dow
--- PCBoard (R) v15.3 (OS/2) 5
---------------