Re: [idn] RE: An idn protocolfor consideration in making therequirements

[C C Magnus Gustavsson <mag@lysator.liu.se>]

On Fri, 11 Feb 2000, ned.freed@innosoft.com wrote:

> Now suppose we ask the DNS operations folks how long it will
> reasonably take for the necessary software to deploy to any
> significant degree, keeping in mind that we're dealing with a
> critical infrastructure component provided by folks who have a lot
> on their plate besides internationalization. I'm not a DNS
> operations person, but I would not be totally amazed if the answer
> that comes back is most conveniently expressed in units of decades.
> 
> So the bottom line is how do you feel about a solution that is purer
> and nicer and has some real attractiveness long term, but which you
> won't see significant benefit from for many, many years? What is
> your choice then?

Upgrade of authorative nameservers:

The software of your own nameservers as well as the nameservers of
your parent zones need to be upgraded. It won't take long before the
software of rootservers and TLD-servers is upgraded. People with
second level domains who delegate third levels domains to others (who
might want international domains) will most likely upgrade pretty fast
as well. (If they don't, I strongly suggest using another third level
domain or a second level domain.)

Upgrade of resolvers/caching nameservers:

This is where it might take some time before software is upgraded. But
not upgrading will not cause trouble for others. If these people want
to be able to benefit from international domains they upgrade. If they
choose not to, well then let them.

Thus, I don't think DNS software is much of a problem with an 8 bit
approach. We will see significant benefit fast enough. I will upgrade
my own resolver software and nameserver software. The nameservers for
domains with international characters and their parent zone
nameservers will be upgraded. That's all that is needed.

About mailservers:

When SMTP is extended to handle international domains, it will most
certainly not be extended in a way that allows e-mail to disappear.

The adopted solution will probably negotiate what protocol to use and
bounce the mail if the server can't handle it. Thus we might have
bounced mail but not disappeared mail. If your mail bounce you simply
have to use an ASCII address instead.

During a transition period of one or a few years few people will want
to use non-ASCII addresses as sender addresses. If mailservers can't
handle non-ASCII domains after that, you either have to choose to use
ASCII only or not communicate with these people. It's not any
different from mailservers which refuse to accept 8 bit e-mails.

/Magnus