Currently, for a fault plane with a dip of 90 degrees we return a dip direction of 0 degrees. This is ok in the sense that it isn't wrong, but not ok in the sense that dip direction is undefined for dip = 90!
Not sure at this stage how we are to provide a dip direction for dip = 90.