I use Star*Net but Carlson has an adjustment routine too. What I like about star net is I don't have to use a point numbers for an observation but simply put a "/" and then the number it is supposed to be. I am going to try to try and use Carlson so I can take advantage of it. I used it for open traverses too. If you double all your angles and close the horizon, you don't have to close your loop to get good answers, but it helps! Your input for systematic errors can be deceiving too, if you have have any redundancies since the program will simply use those to give you a relative error ellipse.