Re: Bogus nic.fr behavior

[Robert Elz <kre@munnari.OZ.AU>]

    Date:        Tue, 6 Feb 2001 15:56:39 +0100 (CET)
    From:        Mats Dufberg <dufberg@nic-se.se>
    Message-ID:  <Pine.BSF.4.30.0102061548380.16896-100000@spider.nic-se.se>

  | Well, we should expect more:

  | * The SOA record must have a working email address.

Do you have any reliable method of testing that, without actually sending
it some e-mail and hoping someone replies, and then associating the reply
with the delegation request that caused it?

I used to like to use SMTP VRFY for this, and tried to get the drums wg
to keep it as a requirement, but too many mailer implementors didn't want
to bother doing a sane implementation (some noise about privacy, but as
you can always use "RCPT To" to test an e-mail address by seeing if mail to
it bounces, that's nonsense - the issue really comes down to whether he
SMTP client is able to figure out what is a valid address, or whether that
gets deferred to later in the e-mail processing).

So, now VRFY can't be relied upon any more - is there some other way
that anyone has found to actually automate this reasonably?

I implement all of your other tests ... plus I make sure that the nameservers
have distinct IP addresses (I'd also like to verify that the distinct addresses
actually reach different servers, rather than just two addresses allocated
to the same server, but again, this one isn't easy to verify).

kre