Toward Understanding Information Models of Fault Localization: Elaborate is Not Always Better