Because (1) an object_id is specific to the address book. A "move" is 
a delete from the old address book and an "add" into the new address 
book. There is no guarantee both are on the same backend, and (2), The 
actual "key" that is used in the group to identify the contact is a 
combination of the address book or share id and the contact's id.

Come to think of it, why is the object_id even changing after moving 
to a different address book?

object_uid only exists in the default SQL schema, while each contact 
from any source must have an ID.

Currently, contacts are stored in the contact list using the 
object_id, which changes after moving to a new address book.

I think that, instead of the object_id, the object_uid should be used, 
as this one does not change.

Regards,

Wim.

Horde Groupware 5.1.1. using Turba.

I have recently been informed that sharing address books is no longer 
an option because the security of the information cannot be 
guaranteed.  I need my address book to be available to several members 
of the organisation immediately and constantly so the only short-term 
option is to export/import and then carry out any subsequent 
amendments in several places (!).  An horrendous and potentially 
dangerous task in itself.  Imagine my horror when I discovered that 
the export/import facility does not maintain the integrity of the many 
contact lists that I have set up.  Each of these now has to be 
recreated manually under each different email address.  If address 
books cannot be shared, the export/import facility has to work.

When moving contacts from one address book to another, any lists 
containing the original contacts should be updated too, if possible.