needed clutches and prints them for the user with 'bsub' commands which
then the user can choose to run or modify

analyses by fetching all the control rules and checking the distributed
conditions that they are all 'DONE'

analysis objects adaptor to fetch the corresponding stats object.
Important to use correct db connection since the analysis objects may be
network distributed across multiple databases

Bio::EnsEMBL::DBSQL::DBAdaptor to match symetry with core,pipeline,compara
This means I can remove all 'hive' adaptors from compara now and the
URLFactory now returns by default a Hive::DBSQL::DBAdaptor.
This also means I can make my graph configuring runnableDBs really
adaptible to run on both Pipeline and Hive systems by doing
if($self->db->isa('Bio::EnsEMBL::Pipeline::DBAdaptor')) {...}
if($self->db->isa('Bio::EnsEMBL::Hive::DBAdaptor')) {...}

when ALL of their conditions are 'DONE')

'living' workers in the hive. Also changed the way adaptors are held
to avoid any problems of the type of DBAdaptor used to create the queen

added method fetch_by_needed_workers for queen to determine new worker
types

so the base adaptor will be a compara DBAdaptor which inherits from the
core DBAdaptor

directory if one isn't specified

a machine.  Very rough and not fully functional.

this logic into the Queen

and summarizes it's state via analysis_stats.

are bound to analysis rather than being distinct

setter analysis share the same adaptor (same database) then use the
logic_name as the url rather than a full url.

added default values of 1 to batch_size and hive_capacity