[provreg] Changes we'd make to EPP, was Re: Domain check in draft-obispo-epp-idn-00.txt

[Keith Gaughan <[e-mail]>]

On 05/01/12 06:49, Patrik Fältström wrote:

> - We would not have the ability for registries to invent their own
> extensions without IETF review (i.e. like other protocols)

An unfortunate consequence of that would be that some registries would just
fork the protocol and use something that's almost-but-not-quite EPP. EURid
went and did this, and when I wrote our EPP client, it was built to account
for nonsense like this, which made the code much more complex than I'd like.

> Example of the 2nd are all the registries that require delegation at time
> of registration (something that should be forbidden, but now some
> registries are like that), that required creation of contact objects and
> host objects before the domain object is created and glued together. If
> at that point in time the create fails, the creation of the other objects
> would have automatically been undone. Maybe as a transaction, I do not
> know. We should have thought about it harder.

Actually, I've no problem with creating contacts, then having the subsequent
registration fail. But what I'd prefer to be different is that I think the
split between hosts as objects and hosts as domain attributes was a mistake.
In addition, I think that if a <domain:create>, &c., request is made and a
set of nameserver hosts are specified on it that have not been explicitly
created with <host:create>, the registry should implicitly create those hosts.

Another thing I'd change is provide a way for registries to signal to
registrars that they've garbage collected contacts (and possibly hosts too).

K.

-- 
Keith Gaughan, Senior Developer
PGP/GPG key ID: 3E896381
Blacknight Internet Solutions Ltd. <http://blacknight.com/>
12A Barrowside Business Park, Carlow, Ireland
Registered in Ireland, Company No.: 370845