Identifiers, Handles and ROID's

[Roger Castillo Cortazar <castillo@nic.mx>]

Hi everybody.
I have a few questions about the globally unique identifiers for objects.

In the reqs. we can read.

3.4.1 Object Identification
...
   [2] Object identifiers MUST be globally unique.
...
   [6] An object's identifier MUST be generated and stored when an object
   is created.

Nothing about the indentifier being returned to the client.

Now, in the EPP specs.

2.7 Object Identification
...
   Globally unique identifiers can help facilitate object information
   sharing between repositories.  A globally unique identifier MUST be
   assigned to every object when the object is created, and the
   identifier MUST be returned within the response for the command that
   created the object.
...

Followed by the algorithm to create such identifier.

But an identifier as described by that algorithm is not returned by the 
<create> command as described in the specs.


And more ....

This could be silly, but I've been reading past year messages on the topic 
of unique handle generation
to try to figure out a few things, and I just may need a few clarifications.

I can see three concepts here:

- The handle as discussed on the list almost a year ago,
- The globally unique identifier as can be read in the drafts,
- The ROID's as returned in the <info> command response.

Are all these the same thing ?

The discussed handle is the local part of the globally unique identifier ?
Are the contact-id and the FQDN for the domains and hosts instances of any 
of these identifiers ?

Greetings.


Roger Castillo Cortazar
NIC-Mexico http://www.nic.mx