Re: [dnsext] New Version Notification for draft-barton-clone-dns-labels-fun-profit-00

[Alex Bligh <[e-mail]>]

--On 7 March 2011 17:33:07 -0800 Doug Barton <[e-mail]> wrote:

> I think you're missing the whole point. :)  Users want to be able to use
> the variant labels (including things like being on the RHS of NS and MX
> records) so the idea is to deal with them on a basis that is as equal as
> possible.

(I've read the draft but am thinking about it before commenting, so don't
take this as a comment on the draft).

I am not sure Nick is missing the point. We do not *know* that users want
to use the variant labels on the RHS of NS and MX records (beyond the
obvious truism that it would be nice to do everything). I am guessing that
most users want to use the variant labels in URLs and email addresses, and
that's it.

Suzanne Woolf is attempting to gather some matrix of requirements and
balance them against disruption. Until such point (that might be never),
all we've got is different proposed solutions with different functionality.
What you're saying is that CLONE (apparently) allows the RHS of an NS
record to be a variant too, which differentiates it from (e.g.) a
CNAME/DNAME combination.

Out of interest, /why/ is a variant being on the RHS of an NS or MX record
useful? This is never visible to end users, and in the majority of use
cases thus far brought up, the RHS of the NS or the MX record gets to be
one user-illegible string beginning "xn--" rather than another illegible
string beginning "xn--".

-- 
Alex Bligh