A possibly stupid random thought: is there a strong barrier in *all* kernels which enforces 127.0.0.0/8 and ::1 to actually *have* to be local?
The 240/4 problem is 5-6 lines of code in *some* UNIX. It wasn't in any sense globally applied. I suspect localhost is somewhat more strongly coded, but I did wonder because Ted's suggestion that use of the literal IP address in either family would the stronger 'keep it local' made me think: what if somebody hand installed a route which somehow took it off-box? I think proscriptive/definitive language over the FQDN/label localhost in DNSSEC is probably still a good thing. IETF is defining behaviours for home.arpa in HOMENET which logistically fall into a very similar bucket (for me at least) so its not like we can't chose to say what behaviours we expect of a label. -G _______________________________________________ DNSOP mailing list DNSOP@ietf.org https://www.ietf.org/mailman/listinfo/dnsop
