On 03/03/2014 01:43 PM, Ted Lemon wrote: > On Mar 3, 2014, at 1:32 PM, Tony Finch <d...@dotat.at> wrote: >> As well as Joe's AS112 argument there is also the question of DNSSEC >> validation - but perhaps we don't want non-DNS names to make any kind of >> sense in this respect... cf. .local > > Indeed, it doesn't make much sense to me that special-use names that are not > intended to be resolved using the DNS should be validateable via DNSSEC. If > they can be validated, it would have to be using whatever protocol is being > used for name resolution (if any). >
+1 This is something I asked about at the app session, and have been wondering; (why) are we worried about non-dns names at all? More importantly, where do we draw the line? "A domain by any other name?" Is anything sequence of characters that happens to maybe contain a dot a domain name? Is it that it *might* end up in some code path that tries to resolve it, even though the normal use doesn't use (the global) DNS at all? To make a crazy example; the left-hand side of an e-mail address also contains dots, but we're not worried about those somehow ending up in a resolution call. Neither were we worried about 'command.com' being resolved (and it does). I'd think that a domain name is only a domain name when whatever protocol it is defined in defines it as a domain name (or whatever undefined protocol uses it in actual dns resolution). What a non-domain name looks like shouldn't matter. Jelte _______________________________________________ DNSOP mailing list DNSOP@ietf.org https://www.ietf.org/mailman/listinfo/dnsop
