University timetabling is a fundamental periodic activity in academic planning. In its general setting this problems is NP-complete. Devising effective strategies for solving it has been a challenge for several decades. Most approach that work in real application either find approximate solutions or only consider constraints of a very specific type. Our approach has been to tackle the full complexity of the problem using concurrent constraint programming techniques. We describe PATHOS, a concurrent constraint object oriented software written in Mozart that has been effectively used in real life situations. PATHOS improves the state of the art in automatic timetabling strategies since it handles problems of a much bigger size than has been considered so far using constraint programming techniques. The stratregies are techniques used in PATHOS are also applicable to other types of planning and scheduling problems.