Why isn there a strict error checking policy in FastGEOs routines?
The way people use the results from computational geometry greatly varies based upon their needs. In some situations the input data is known to be of a certain type, hence additional checking for that particular type is not needed and would cause a great deal of inefficiency in the code. An example of this situation is in the calculation of a circumcenter based on three unique points. The base requirements for the routine are that the three points used as input not be collinear to each other. The routine assumes the points passed are not collinear and completes its calculations with that assumption in mind (there is still minor division by zero error checking done, which relates to the points being collinear). However in a situation where you can’t be sure that the three points will be unique and will not be collinear to each other error-checking must be done manually. The amount of error checking you do will determine how robust your sequence of routines will be and in the end how acc
