> On 26 Mar 2019, at 18:23, Brian Dickson <brian.peter.dick...@gmail.com> wrote: > > The options are, new RRtypes that require resolver upgrades, or RRtypes that > are handled by the client application (browser), which benefit from (but do > not require) resolver upgrades.
The current draft is neither of those (and I think maybe you excluded it from your list because you don’t consider it viable, for good reasons, but...) Whatever new thing we come up with, my view is that it will need sibling address records for backwards compatibility, otherwise no-one can deploy it. (Or at least, that new thing won’t make my job easier any time soon.) So there has to be some kind of provisioning or authoritative hack to make it easy to automatically add address records that have the same effect as the ANAME or HTTP record that they are supporting. (Remember, the existing non-standard alias stuff exists because it is easier than standard DNS.) Ray Bellis thinks (quite reasonably) that there is a good chance that an HTTP specific record can be deployed fairly quickly, because browser upgrades are impressively fast these days. I’m less enthusiastic about telling DNS admins to look at caniuse.com before they decide whether to put in an HTTP record. (On the gripping hand, based on the support queries I have answered, an HTTP302 record [an http redirect in the dns] would be a lot closer to what many webmasters expect!) Tony. -- f.anthony.n.finch <d...@dotat.at> http://dotat.at _______________________________________________ DNSOP mailing list DNSOP@ietf.org https://www.ietf.org/mailman/listinfo/dnsop