The Last Slide
Can successfully use ODBMS for complex HEP data
Objectivity “database federations” a good match with LHC data storage needs
Globally distributing the database will bring special problems for analysis
Still struggling with a worthy replacement for PAW: OO Analysis tools in their infancy
LHC++ not yet proven usable for end user physicists
PAW->++ (really Fortran->++/Java) inertia
Java very promising for Event display and some analysis