Stats-QC : be able to open really old studies

Seen on Lille Server : 500 error in this case... It may only be a problem with initialization of diversity variables in this case (should be initialized outside the if, may be initialized before the bigger if)