On 8/15/06, Andreas Hartmann <[EMAIL PROTECTED]> wrote:
IMO the UUID should be orthogonal to the publication which the
document belongs to. The UUID is forever, even when you move the
document to a different publication.
I think the most reasonable way is to use a common storage for all
documents, regardless of the publication they belong to.
In Domino, within a Database each Document has a ReplicaID used to
synchronize the Documents in all Replicas of the Database. If a new
Replica (exact copy that can be synchronized) of the Database is
created, the ReplicaIDs are not changed. If a new Copy of the
Database is created, all the Documents are assigned a new ReplicaID.
If the Document is copied to another Database, it is assigned a new
ReplicaID. If the Document can be copied to the same Database, the
copy is assigned a new ReplicaID.
For Lenya terms, replace Database with Publication, and ReplicaID with
UUID. Do we have a term for a Publication that stays synchronized
with another Publication?
If a Document retains its UUID when it is copied to a different
Publication, then the two Documents should contain the same
information. This requires frequent communication between the
Publications. Does Lenya have that ability?
If copies of a Document will not remain synchronized, the UUID should
be changed during the copy operation.
---
In my last post, I suggested that containing Documents for multiple
Publications in one datastore would be easy. That implies multiple
Publications can contain the same Document. But they would share the
same physical instance of the Document; if it is changed, all the
Publications would show the change.
If the Document is Copied within a Publication, it must be assigned a
new UUID, or there will be conflicts. If multiple Publications are
sharing the datastore, and the Document is Copied from one Publication
to another, the Document must be assigned a new UUID. The two copies
share everything up to the point of Copying, but then they are
maintained separately, and must have their own unique identifiers.
---
Joern,
Each UUID is supposed to be unique in the world. There are enough of
them that we cannot exhaust them. Of course, their uniqueness depends
on the generators. If two generators start with the same number and
increment sequentially, they will produce duplicates. All modern
generators are supposed to generate UUIDs that are part random, part
time-based, and part based on the source of the UUID. There are a
finite number of them, and statistically there must be a collision
someday, but it is expected the universe will end before a meaningful
collision occurs.
solprovider
---------------------------------------------------------------------
To unsubscribe, e-mail: [EMAIL PROTECTED]
For additional commands, e-mail: [EMAIL PROTECTED]
