Re: [idn] Which names are valid?

[Erik Nordmark <Erik.Nordmark@eng.sun.com>]

> The definitions in that draft are not entirely satisfactory. For example,
> consider "_ldap._tcp.π.com". This is clearly a domain name, and it is
> clearly internationalized, so the most sensible thing to call it would be
> an internationalized domain name. However, it is not a hostname (although
> "π.com" is). It makes no sense to lose the useful distinction between
> hostnames and domain names just because they are internationalized.

So is "_ldap._tcp.π.com"" a domain name with a hostname component??
[my mail UA seems to have problems with the characters above - it might not
look the way you sent it.]

The reason we can't just drop the distinction between host names and domain
names is that many application protocols handle host names under the
restrictive specification of valid host names. When looking at transition
issues the impact on those application protocols is key.

So should we say something specific (e.g. as part of the scope
in the requirements document) about domain names with
host name components or substrings?

   Erik