From: John Palmer <jpalmer@bl...> - 2001-10-25 16:11:30
I fully agree... registry level assignment works nicely...
At 09:03 AM 10/25/2001, you wrote:
> > -----Original Message-----
> > From: Daniel Manley [mailto:dmanley@...]
> > Sent: Thursday, October 25, 2001 11:18 AM
> > To: wessorh@...
> > Cc: 'epp-rtk-devel@...'
> > Subject: [Epp-rtk-devel] more detailed differences between
> > EPP 02 and 04
> > The remainder of the changes have to do with contact objects. In EPP
> > 02, the registrar would supply only contact information (name, org,
> > address, etc...) for a contact create request. The registry would
> > return the new contact object's ROID. This would be the object
> > identifier in further requests regarding that object (info, update,
> > transfer, delete). Plus, this ROID would be used to associate the
> > contact with domains. When creating a brand new contact, the check
> > command would have no use since the registrar has no control over the
> > contact's identifier.
> > In EPP 04, the registrar must supply the contact object's identifier.
> > Like a domain's and a host's name, the contact ID cannot
> > already exist
> > in the registry for any registrar. The registrar could
> > typically use a
> > DB sequence, perhaps in combination with a short form of the
> > registrar's
> > name to prevent collisions with other registrars. So here,
> > the contact
> > check command becomes very useful to avoid failed create
> > commands due to
> > duplicate IDs.
> i'd like to ask why this change was made.
> i thought it was better for the registry to return a unique identifier for
>the contact handles that way there could NEVER be any type of conflict
>Epp-rtk-devel mailing list