Re: [DNSOP] discussion for draft-appelbaum-dnsop-onion-tld-00.txt

[Andrew Sullivan <[e-mail]>]

Dear colleagues,

On Mon, Mar 16, 2015 at 10:16:37PM +0000, Jacob Appelbaum wrote:

> I realized after uploading that I hadn't sent this along for discussion.

> > Name:		draft-appelbaum-dnsop-onion-tld

I've read this draft.  I have a few comments.

To begin with, in general I think this document is on the right path
and something very close to it should be published.  It's
narrowly-focussed, I _think_ it meets the template requirements of RFC
6761 (but see a remark below).  It's clearly a technical distinction,
because this approach is a backward-compatible user interface
namespace without any real DNS vestige at all: it's not intended to be
a general-purpose identifier system on the Internet, but rather as I
understand it a cryptographic identifier in the Tor network.

I have a couple nits:

In section 1, it'd probably be useful to pick up the term "domain name
slot" that was introduced by RFC 5890.  This is not an i18n case, of
course, but that's sort of the point: this is an example of something
that more or less fits in places people already use ordinary DNS
domain names, but this name is special and actually not in the DNS as
it shows up in the domain name slot.

In section 4, 3-5, what if a "synthetic" NXDOMAIN gets generated and
cached?  Will that have any effect on .onion resolution?  If this is
explained in detail in some thing I've failed to follow, a simple
reference would be enough.

Best regards,

A

-- 
Andrew Sullivan
[e-mail]