Based on well understood track parameter errors the natural next step is to make vertex fit based on Kalman filter techniques. The fit should be done for primary and secondary vertexes.
Reassess how much FORTRAN code use tables. Move table definition from idl to h files.
Get rid of stic