Re: [DNSOP] Comments on draft-ietf-dnsop-alt-tld-05

"John Levine" <> Sun, 09 October 2016 17:36 UTC

Return-Path: <>
Received: from localhost (localhost []) by (Postfix) with ESMTP id 3DA91129594 for <>; Sun, 9 Oct 2016 10:36:59 -0700 (PDT)
X-Virus-Scanned: amavisd-new at
X-Spam-Flag: NO
X-Spam-Score: -1.9
X-Spam-Status: No, score=-1.9 tagged_above=-999 required=5 tests=[BAYES_00=-1.9, LOTS_OF_MONEY=0.001, RCVD_IN_DNSWL_NONE=-0.0001, SPF_PASS=-0.001] autolearn=ham autolearn_force=no
Received: from ([]) by localhost ( []) (amavisd-new, port 10024) with ESMTP id EjLZJRJY5m1r for <>; Sun, 9 Oct 2016 10:36:57 -0700 (PDT)
Received: from ( [IPv6:2001:470:1f06:1126::2]) (using TLSv1 with cipher DHE-RSA-CAMELLIA256-SHA (256/256 bits)) (No client certificate requested) by (Postfix) with ESMTPS id 6CF9B12958D for <>; Sun, 9 Oct 2016 10:36:57 -0700 (PDT)
Received: (qmail 55284 invoked from network); 9 Oct 2016 17:36:55 -0000
Received: from unknown ( by with QMQP; 9 Oct 2016 17:36:55 -0000
Date: 9 Oct 2016 17:36:34 -0000
Message-ID: <20161009173634.34142.qmail@ary.lan>
From: "John Levine" <>
In-Reply-To: <>
X-Headerized: yes
Mime-Version: 1.0
Content-type: text/plain; charset=utf-8
Content-transfer-encoding: 8bit
Archived-At: <>
Subject: Re: [DNSOP] Comments on draft-ietf-dnsop-alt-tld-05
X-Mailman-Version: 2.1.17
Precedence: list
List-Id: IETF DNSOP WG mailing list <>
List-Unsubscribe: <>, <>
List-Archive: <>
List-Post: <>
List-Help: <>
List-Subscribe: <>, <>
X-List-Received-Date: Sun, 09 Oct 2016 17:36:59 -0000

>1)  applications may query for "alt." for some RR. If the
>    resolver is not "alt-tld aware", or alt-tld functionality
>    has been disabled, then the request will return NXDOMAIN;
>    if instead the mechanism is supported, it would return
>    NOERROR with an empty answer. Note that this is backwards
>    compatible with non-alt-aware implementations since they
>    would leak in DNS and NXDOMAIN will be returned by some
>    other server.


>2b) as above, but it might be specified that the query is
>     to be done on a TXT RR specifically and that, instead
>     of the empty answer, a record might be returned with
>     contents that can be used to in some way ascertain that
>     the handler is actually for the implementation we expect
>     it to be (for example a software name or version, or
>     protocol "UUID"?)

I fear we are tumbling down the slippery slope into a tar pit here.

There is absolutely no standardization or commonality in the protocol
switches that people use for non-DNS names.  For .local it's in
library routines like gethostbyname() that map names to IP addresses,
for .onion it's in a SOCKS proxy that opens a connection to a remote
host and for those of us old enough to remember .uucp and .bitnet, it
was in a switch table in the mail configuration that determined what
program to call to do the next hop.  It's not much use providing an IP
address lookup handler for something that doesn't use IP addresses.

I hate to bring this up, but we might want to reconsider whether .alt
is still the best name for this hack.  For us old nerds, .alt reminds
us of usenet and uucp of yore, but in American politics it's some
unpleasant fringe groups with which I would not want to be associated
even a little bit.  Keeping in mind that we want to encourage people
to use .alt, I'd rather find something with no political (or other)

I would suggest .app except that Google paid ICANN $25M for it last year.