Re: [DNSOP] Last Call: <draft-ietf-dnsop-onion-tld-00.txt> (The .onion Special-Use Domain Name) to Proposed Standard

[Jacob Appelbaum <[e-mail]>]

On 8/7/15, Edward Lewis <[e-mail]> wrote:
> On 8/7/15, 10:29, "DNSOP on behalf of Wendy Seltzer"
> <[e-mail] on behalf of [e-mail]> wrote:
>
>>You might find https://spec.torproject.org/ helpful as a listing of
>>various tor specs and design documents, if you prefer that to a git
>>repository.
>
> That's the site I've been using.
>

What do you specifically wish that it had? Doesn't the git repository
provide you with specific versioned hashes for any given document?

>
>>
>>While Tor has not necessarily used IETF conventions, the project has
>>long been committed to public documentation of its design and protocol
>>choices. Tor distinguishes between "proposals," not yet implemented, and
>>specs.[1]
>
> It's fine to "not use IETF conventions".  (One of my ratholes is that I
> find researching IETF documents rather frustrating - just to underscore I
> am not saying Tor's documentation is deficient compared to IETF standards.)
>
> Still, the documents I have access to do not give me a deep enough sense
> of, well, why the names are different from DNS domain names.  I presume
> they are from the email discussion, but what I am reading in the documents
> - and I stress "reading in the documents" meaning that might be the gap -
> doesn't give me enough background.

This is clearly documented in the specifications - including the
lookup process and how the .onion names are constructed.

I'm really confused as to how your question of "why the names are
different from DNS domain names" isn't answered by Tor's rend-spec?
Should we make an explicit reference to Zooko's triangle or something?
o_0

>
> As far as stability of the documents, referring to a document by URL only
> (which is accepted in IETF documents at times) isn't generally accepted.
> I admit this is a bit of a red herring point, because this can be changed,
> but if there were other means to refer to the document in a reference
> citation, it would help.
>

Do you mean to suggest that https://spec.torproject.org/rend-spec
isn't enough? Wouldn't
https://gitweb.torproject.org/torspec.git/tree/rend-spec.txt?id=69f0faec555d3c7d627aa1de73dafee8f8d63f84
be specific enough?

> E.g., randomly typing a four digit RFC number:
>
> http://www.rfc-editor.org/rfc/rfc7392.txt
>
> vs.
>
> Dutta, P., Bocci, M., and L. Martini, "Explicit Path Routing for Dynamic
> Multi-Segment Pseudowires", RFC 7392, DOI 10.17487/RFC7392, December 2014,
> <http://www.rfc-editor.org/info/rfc7392>.
>
>
> That is, full title, RFC document identity, date and authors/editors (not
> in that order).
>

The git repo shows authorship down to the line number as well.
Documents not specifically authored by named persons or teams are
authored by the individuals in the git commit and/or the Tor Project
inc., I believe.

> PS - I don't mean to harp on this.  I'd hoped to have someone send me
> links to other documents because I want to learn more about the names in
> ".onion" and other identifiers in Tor.
>


Please see:

  https://gitweb.torproject.org/torspec.git/tree/rend-spec.txt?id=69f0faec555d3c7d627aa1de73dafee8f8d63f84

As well as:

  https://gitweb.torproject.org/torspec.git/tree/proposals/155-four-hidden-service-improvements.txt?id=69f0faec555d3c7d627aa1de73dafee8f8d63f84
  https://gitweb.torproject.org/torspec.git/tree/path-spec.txt?id=69f0faec555d3c7d627aa1de73dafee8f8d63f84#n36
  https://gitweb.torproject.org/torspec.git/tree/address-spec.txt?id=69f0faec555d3c7d627aa1de73dafee8f8d63f84#n48

And more generally:

  https://gitweb.torproject.org/torspec.git/tree/glossary.txt?id=69f0faec555d3c7d627aa1de73dafee8f8d63f84
  https://gitweb.torproject.org/torspec.git/tree/proposals?id=69f0faec555d3c7d627aa1de73dafee8f8d63f84

Specific implementations are probably also useful. Tor's C
implementation and Orchid (Java Tor implementation) are probably easy
enough to study.

Happy to answer other questions as well...

All the best,
Jacob