[Date Prev][Date Next][Thread Prev][Thread Next][Date Index][Thread Index]
Re: [idn] host name vs. domain name
--On Friday, 17 March, 2000 15:02 -0800 "David R. Conrad" <David.Conrad@nominum.com> wrote:
> Another aspect of this discussion is the whole question of the in-addr.arpa
> tree. By using UTF-8 names, it becomes "a bit" more challenging to try to
> interpret log files, traceroute output, etc. I am not sure this is of
> critical importance, but it is something that should be taken into
Of course, it also implies that, if special versions of the name server or resolver, or interfaces to them, have to be deployed to make i18n machinery work properly in practice, we are in a different
sort of deployment trouble. An enterprise in an ASCII-using country might postpone or avoid upgrading servers, resolvers, and applications if it concluded that there would be no legitmate business
reason for international names to be used. But the need to deal with reverse-mappings implies that it might be under some extra pressure to upgrade (good news in some respects, bad in others). For
example, MTAs often reverse-map addresses as a sanity check on incoming traffic. If the values that come back from the resolver get the MTA into trouble, the temptation would be to filter and reject
all SMTP connections originating from hosts that exhibit non-ASCII text in reverse mappings.
And, of course, reverse mapping issues make theories based on having two names for many hosts -- one 10646 (encoded in UTF-8 or otherwise) and one in ASCII -- a bit more complicated.