Re: interpretation of 'EPP idempotency'

[Thomas Corte <Thomas.Corte@knipp.de>]

Hello Daniel,

On Tue, 18 Sep 2001, Daniel Manley wrote:

> Thomas,
> Would this example be pulled from the operation of the .info registry?
>  They are using a older version of the spec.  In that release, the
> contact create command contains no unique identifier from the registrar,
> so multiple requests for the same contact will create duplicated in the
> registry.  In the current version of the spec, the registrar must send a
> registry-unique identifier for the contact.  If the registrar uses a
> consistent method of generating a unique id (maybe a hash of the name
> and email address?), then multiple creates of the same contact would
> fail because the id won't be unique after the first create.

In fact, you are right - I deduced this example from Afilias'
current practice for contact creations and wasn't aware that they
use an old spec version.

However, I would highly appreciate a *uniform* solution for error
recovery, which is not present even in the most recent EPP draft (as
Klaus' recent mail pointed out).

Regards,
_____________________________________________________________________
     |       |
     | knipp |                   Knipp  Medien und Kommunikation GmbH
      -------                           Technologiepark
                                        Martin-Schmeisser-Weg 9
                                        D-44227 Dortmund
     Dipl.-Inform. Thomas Corte         Fon: +49-231-9703-0
     Thomas.Corte@knipp.de              Fax: +49-231-9703-200