I had imagined, in that simpler time, that users never saw alt, but imported other-name-space calls mapping into gethostbyname() acquired .alt unless an override was present. Ie, it was a more formal label of inclusion of a non-DNS namespace so the alt label was in-system, not visible in the other namespace.
This is how I recall the mapping from the pre-USENET news systems like the one Jacob Palme ran, on Dec-10s was intruded into usenet news: his stuff had a qualified form which was legal in-system in usenet, but was acquiring a magic label on entry from the Dec-10 world. Lots of sendmail.cf files did transforms which added/removed/appended/re-ordered mail addys to get magic tokens in and out. -G On Thu, Feb 12, 2015 at 10:49 PM, Ted Lemon <ted.le...@nominum.com> wrote: > On Feb 12, 2015, at 1:44 AM, George Michaelson <g...@algebras.org> wrote: > > Technology wise, this is short, and simple and clear. Would we had this > before .onion eventuated, and dare I say it even .local from another time > and place. WIring a TLD to be used for alternate namespaces so that we can > safely anchor non-DNS names into the DNS and avoid repetitious stupidity is > a good plan. > > I think that if .alt had existed when .local was defined, nothing would > have been different. From a UI perspective, using an extra label to > identify that a particular domain is special-use is worse, because it's > more typing, and doesn't really make sense--which makes more sense: > "myhost.local" or "myhost.local.alt"? > > So I think this work is worth doing, because there are probably cases > where .alt will do just fine, but I don't think it gets rid of the problem > of top-level special-use domains. The "SHOULD" recommendation in the > document is the right level of applicability: "MUST" would be wrong, and I > appreciate that the authors did not use it. > >
_______________________________________________ DNSOP mailing list DNSOP@ietf.org https://www.ietf.org/mailman/listinfo/dnsop
