Re: [DNSOP] Root reasons (aka "why") - HTTP vs SRV vs ANAME vs CNAME vs URI vs NAPTR

[Brian Dickson <[e-mail]>]

Patrik Fältström wrote:

> Note changed subject...
>
> [rest of message cut from reply]

There is a major semantic difference in the NAPTR/URI RDATA and how/where
it is handled, which is at the HTTP application layer (i.e. 3xx rewriting).

This differs from the other 4 RRTYPEs, where only the A/AAAA values are
sought by the client (i.e. at the transport layer).

It boils down to what is in the browser bar/certificate, and which meaning
of "canonical" is desired (maybe "correct" isn't the right term in this
case).

Given that this is something that we're trying to use to replace CNAME, the
smallest degree of semantic change is probably likely to be the least
contentious.

IMHO, those two are pretty much out of scope, at least for this discussion.
(I don't see a problem advocating for their use, but those are existing
types, and nothing stops anyone from using them.)

No offense intended.

Brian