On 11/09/2020 18:39, Paul Hoffman wrote:
Plus, why couldn't this be extended in the future to recursive nameservers as
well? In which case, the name should be resolver genre agnostic.
There is already a draft for RESINFO. The difference is important, because a
single host might be both a resolver and an authoritative server.
I don't believe two separate types are needed. As you return formatted
data, information on the type (recursive vs authoritative) can certainly
be in the data and there should probably be enough space in that data to
store both cases, as needed.
Maybe something like "CAPABILITIES" or "ABILITIES"?
(or shorter version)
And you're thinking those aren't used in other protocols? :-)
The difference probably being on the fact that EPP is also very tied to
the same area, aka domain names.
So the idea is not just to use a unique name across the whole word, but
at least not a name already extensively used in another protocol in the
same "world" of domain names.
That was just my suggestion (do not reuse authinfo name), feel free to
ignore.
(but your email triggered that reflex in me as I have seen, in the realm
of just EPP, so much abuse or misunderstanding of the authInfo term,
that I wanted to convey the idea that even more confusion could be a curse).
--
Patrick Mevzek
_______________________________________________
DNSOP mailing list
DNSOP@ietf.org
https://www.ietf.org/mailman/listinfo/dnsop
