ODB no longer uses DPL layer of Berkeley DB, persistent classes have to
implement the Serializable interface, annotations are no longer used,
use transient modifier for variables that should not be persistent.
Entity reference errors will no longer occur, however it is better to
use ids if the entity that is being referenced is stored in the db too!
(or you will have two different objects). DB is cached for 5 minutes
then it will write its log/cache to disk, transactions are no longer
used.
Transformed apparent non-entity-connected object-relational dependencies
into looser coupled ID-keyed relations to ensure referential integrity
Added null pointer checks
Removal of null elements out of survey tool collection
Shouldn't create a new objectId each server run, by using a duplicateId
database that takes other things (such as planetcrc, templatecrc,
building, container, cell, coordinates)
Currently best solution I can see that doesn't consume a brand new lot
of 20,000 objectids each bootup.
Light2