Review what to do with the following three areas of POOL:
- POOL-ORA (uses CORAL and was used by CMS; now not used, but I understand from DavidM that this could be interesting for ATLAS?)
- POOL collections (uses CORAL and is used only by ATLAS)
- POOL/ROOT (does not use CORAL and is used by ATLAS and LHCb).
About CORAL, the overlap between experiments is much larger than for POOL. ATLAS and LHCb both use CORAL through COOL; CMS also uses it for conditions; in addition ATLAS uses CORAL for the detector geometry, trigger configuration and through POOL collections; and there are possibly other uses I am forgetting. Most of CORAL (eg Oracle, SQLite, Frontier, XML...) is used by at least two experiments and often by all three; the only parts that are presently used only by one experiment are the LFC replica service (LHCb) and CORAL server (only ATLAS HLT so far, though this could be useful for LHCb too at some point).
About COOL, several features are only used by ATLAS and not by LHCb - but it would be very hard to separate them out as they are just additional methods of the same interfaces. Note also that both ATLAS and LHCb have their own 'IovDbSvc'-like component to read data from COOL into Athena/Gaudi, with some commonality and (I think) several experiment-specific features.
