On 2009-12-25, at 06:02, Arnt Gulbrandsen wrote: > What is the actual difference between the proposed sink.arpa and the existing .invalid? (a) Our idea when we chose that name was to try and make the policy environment within which the (non-) assignment rule was to be instituted clear. The administration of ARPA is fairly clearly defined, and lies fairly clearly within the policy control of the IETF and the IAB. The administration of the root zone has a far greater audience of participation, and is hence more likely to be subject to future change. Naming the (non-existent) name under ARPA avoided this potential headache. (b) SINK.ARPA is a hostname whereas INVALID is not, according to certain interpretations of (I think) RFC 952. (I think I last saw reference to this when Mark Andrews was discussing potential problems with apex MX records in TLD zones, in the context of discussing expansion of the root zone). This perhaps makes INVALID less appropriate for applications where a non-existent hostname is required. Joe _______________________________________________ Ietf mailing list Ietf@xxxxxxxx https://www.ietf.org/mailman/listinfo/ietf