On Fri, Nov 1, 2019, at 15:00, Alexander Mayrhofer wrote:
> Hello everyone,
>
> i've create a rough cut of a draft describing an EPP extension to
> convey Domain Suggestions. I'm planning to do a small side meeting
> with interested people outside to gauge interest, but i'm also happy
> to grab 5 minutes off the regext session in Singapore to present the
> idea.
>
> As always - feedback highly appreciated!
Some others more substantive comments that I think have huge consequences:
- generating suggestions is surely costly for the registry, and hence should
not be done if not strictly necessary; hence said otherwise, besides
the negotiation at login, the registrar should or must have a way to say, per
command
"i want a domain:check with suggestions" or "i want a domain:check without
suggestions".
Like in the launchphase extension, where a single domain:check can return
different
results depending on the phase.
Besides being costly to generate by the registry,
the registrar has no way in advance to know the size of
the response which can be a big problem for it.
This may show in fact that extending domain:check is maybe not the right way
to do and a specific separate command may be better
- multiple registries handle multiple TLDs: the registrar asking for suggestions
may wish to restrict the TLDs of suggestions, and hence need a way to convey
that.
What if the registrar is not accredited for all TLDs in that registry?
Extending that, registrars may want to have more knobs to have the registry
returns results tailored to their needs:
* domain length
* using IDNs or not
* "language" or geographic target
* use of hyphens or not
etc.
I am sure everyone can see dozens of possible such settings.
Another way to not explode in complexity but still enable various results,
the registrar could pass a "profile" name, which would be any opaque id
but the registry would announce, out of band, that profile named X means search
is done using such and such cases, etc.
It is not clear also if domain names returned are guaranteed (at that time)
to be available for registration or not.
At this stage, to be honest, I fear writing a specification is too early.
Various people have expressed interests, there seems to be already some
private/local implementations on various protocols,
and even a past EPP extension, so I think for now it is maybe better to be
collecting
the use cases, the goals wanted, and looking at what exists,
to be able to first define what work needs to be done... before doing that work.
I guess I am looking more at past examples like
RFC 3375 "Generic Registry-Registrar Protocol Requirements"
or
RFC 7485 "Inventory and Analysis of WHOIS Registration Objects"
that were written before any work took place.
--
Patrick Mevzek
p...@dotandco.com
_______________________________________________
regext mailing list
regext@ietf.org
https://www.ietf.org/mailman/listinfo/regext
