Minutes for HRPC at IETF-94
minutes-94-hrpc-1

2015-11-03 14:36:24+0900
------------------------

HRPC at IETF-94



- Beginning (5 min)
        Agenda Bashing
        Jabber scribe, note takers
          - dkg note taking
          - Melinda Shore jabber scribing
        Notewell
        Introduction
- - Status of proposed research group (5 min)
- - Context of research (10 Min)

 -- Scott Bradner objects to the white-on-red as a violation of
    usability guidelines ("speaking of human rights")


- - Discussion of 'A Case Study of Coding Rights' (20 min)
https://mailarchive.ietf.org/arch/attach/hrpc/pdfbyB1Dp.pdf

 -- Corrine Cath speaks remotely
 
   carnivore
   post-Snowden pervasive monitoring
   OPES
   Middleboxes
   Status 451

when does the IETF encode values:  needs three points

 clear technical reason
 no economic or political pressure against it
 problems directly impact core goals (openness, etc)

recommendations:

 * find ways to have HR guide protocol development.
 
 * more participants who are direct custodians of human rights within the IETF

 * emphasize importance of the four key principles from Clark et al.

Linus Gasser -- i think it's really brave to try to put HR into
                protocols at all, but others understand human rights
                differently.  How should we engage with people who
                have a different backround toward human rights?

Corrine Cath -- some people do say there is no universality, HR is a
                very western notion.  but i've yet to encounter a
                viable alternative.  I'm assuming we want a world
                where freedom of expression is maintained, and the
                UNDHR is the best we've got at the moment.

Robin Wilton -- the concern that there be no political opposition is
                really troubling.  there's always going to be someone
                who wants to not do it.  Sometimes we just need to
                push back.

Corrine Cath -- i'm not saying it shouldn't, i'm saying if there is a
                lot of political resistance, it *won't* be done.  How
                do we ensure that the IETF is aligned with the UNHDR
                in a way that certain actors don't walk away from the
                table.

Jonne Soininen -- If personal morals were encoded, i'd say the
            reverse: the IETF has always been about technically sane
            solutions.  the e2e argument has been something that we've
            used because it makes technical sense.  many IETFers have
            thought "When you do it right technically, you insert HR
            automatically" -have you thought about it that way around?

Corrine Cath -- i havne't thought about it that way.  separating
                personal views from what you do is not something we
                can do effectively.

Ted Hardy -- to answer Jonas: when you talk about "techincally right",
             you're encoding your assumption about what is right.  An
             example is a catenet instead of an internet (catenet has
             firmer borders between networks; internet is that each
             node is part of the same overlay).  there are values for
             catenets instead of internets, but these choices are
             value choices.  You can model it as freedom of
             expression, but i see it as more about freedom of
             association.  "the internet is for anyone" means we
             should be open to interconnection to all parties.  This
             is encoding technical values.  name-drops someone fancy
             which i missed.

Jonne Soininen -- not every engineer thinks about this as the human rights
            principle.  this is chicken and egg.  Was e2e originally
            an HR argument, or was it a technical model?  which came
            first?

Corrine Cath -- Baran wanted a network that could survive nuclear
                attack.  That's a particular view of what network you
                need.

Robin Wilton -- technical and ethical dimensions: Sarah S?? from
                University of Vienna makes clear argument that there
                is no ethically-neutral design.

Ted Hardy -- As to what came first: what i see now is the aim that
             everyone has the ability to join the network.  it doesn't
             prevent that and it should scale as much as possible.
             Whether we meant that as "the right to connect"
             vs. "striving for the ability to connect" has the same
             effect.

Jonne Soininen  -- now we take ethics and openness more seriously, but maybe
             that wasn't always the driving force.




- - Discussion of Methodology draft  (30 min)
https://tools.ietf.org/html/draft-varon-hrpc-methodology-01
                Cases
                        IP
                        DNS
                        HTTP
                        P2P
                        XMPP
                        VPN
                Definitions
                Way forward


- - Discussion of Glossary draft	(10 min)
https://tools.ietf.org/html/draft-dkg-hrpc-glossary-01

Stephane Bortzmeyer -- how do we organize how to make fixes on these
                       documents?  XMPP writeup has false statements.
                       How are we supposed to organize?  should we use
                       issue tracker?

Avri & Niels -- changes have not been deliberately added.

John Levine -- your concentrating on Section 19, but security of the
               person is also relevant.  f. ex. sxsw gamergate
               horrorshow is relevant.  why isn't it in the draft?
               It's always FreedEx

Avri Doria -- we tried to start with just one, to avoid boiling the
              ocean, but i agree those are critical ones to deal with.
              We should look at the proposed charter to consider more
              than just what we've started with.

John Levine -- i'll talk with you later.

Jeff Jack (?) -- when i look at accessibility, i see it as a human
                   right in itself, but you've got it as a piece of
                   "right to political participation"

Niels -- we're comparing from UNHDR over to technical terms.  that's
         why some terms are on the left, and some on the right.

Stephane Bortzmeyer -- we have more subtle points, but hrpc@irtf.org
                       is too quiet.  f. ex. I challenged comments
                       about IP mobility, but no one responded to my
                       challenge.  We need more discussion.

Avri -- methodology document is just a starting point, i agree we need
        more discussion.

how many have read methodology --
  show of hands is very small

Avri -- that's why there's not enough discussion.

Bob ?? -- these rights conflict with each other.  Just picking a few
          isn't helpful because the interesting stuff is in the
          conflicts.

Niels -- agreed, these are interdependent and interrelated and need to
         be balanced, but before we do that we need to define them.

Avri -- I don't think the RG is going to solve the problem.  hopefully
        it's going to help us understand it better instead.

Robin Wilton -- rephrasing Niels and Avri: i don't think the work of
                the group is about defining these rights.  i tihnk
                it's about trying to understand what these rights
                involve and how you implement it.  Bob's right, it's
                not a neat hierarchy.  but that shouldn't stop us from
                looking at them one by one.

Nalini Alkins -- i've been looking at how censorship attacks take
                 place. 
                



- - Discussion of report draft (10 min)
https://tools.ietf.org/html/draft-doria-hrpc-report-00




- - Discussion of 'The Internet is for End Users' draft (20 min)
https://tools.ietf.org/html/draft-nottingham-for-the-users-02


mnot displays W3C priority of constituencies

surfacing the constitutencies makes it a bit easier to think about the
arguments you run into.

failing end users means failing the goals of the internet


John Levine -- i read the draft, and my initial response was "duh,
               this is obvious".  My only concern is that we're
               notoriously bad about knowing what is best for our
               users.

mnot -- right, this draft is in some ways a very small step.  it's
        just sort of a stake in the ground we can use against OPES

Peter Koch -- priority of constituencies is self-referentially invalid
              because it is a theoretical principle that says
              theoretical principles should go last.

              my users might not be your users -- couldn't the
              principles be used as a tool to [missed]

mnot -- "end user" is not usually something that refers to router
          administrators.

Andrew Sullivan -- the term "end user" has the ability to hide a lot
                   of sins.  we don't know who our end users are in
                   many cases.  Suppose HomeNet coalesced quickly
                   around some routing protocol that made it easy to
                   go in and support the system remotely on the part
                   of the ISP.

                   There are two end users: the owner of the home, and
                   the guy stuck at the helpdesk who has a claim of
                   being a legit end user of this protocol.

                   The W3C use cases involve a human.  IETF work often
                   is just machines talking to each other.  I don't
                   think that fascists come to the IETF.  I think the
                   issue might be about focusing on "real problems".
                   can we re-cast this in terms of "how do you
                   identify a real problem"?

                   Using abstract properties instead of end users
                   might be better.

mnot -- yes, a toolbox of sticks to bet down bad proposals would be useful

Ted Hardy -- instead of trying to abstract this down, we consider this
             in the context where end users are effected.  Maybe some
             drafts have no effect on end users, so they have a
             section that says "no issues", just like when IANA isn't
             impacted by a draft.

Lee Howard -- I like the short draft!  I agree that end user depends
              on the thing you're designing.  I want to ask: "better"
              could mean "simpler" or it could mean "more knobs to
              turn"

mnot -- i agree that the relevance to the user might be less at higher
        layers, but lower layers do also have effects on end users.

Ted Hardy -- agree.

Robin Wilton -- leave "end user", since we all understand it's a
                shorthand.  I think we need to understand "end user"
                as "those affected".  f. ex. wrt privacy, my privacy
                is affected by protocols used by other users.

Jeff Jack -- i'm surprised that the IETF has no understanding of
             what's best for end users given what we've just talked
             about with the UNDHR.  these don't compute.

Bob Hinden -- we have insight, we should be able to use it, don't say
              we have no specific insight.

dkg -- the statement is saying that we as the IETF have no specific
       insight into what's good for everyone, because the world is
       more diverse than this room.  The IETF needs to become more
       diverse.

Bryan Ford -- our insight as a group of geeks is anything but a
              representative sample of the internet user community.
              Maybe we can solve that?  Maybe we need a democratic
              root of trust for the Internet.  Who has heard of the
              deliberative poll process?  can we get an ordinary group
              of users to come to the IETF and ask them questions?

Niels -- bring it to the mailing list.







- - Open discussion other drafts, papers, ideas (20 min)
   no takers
- - Charter update status (if necessary - 5 Min)

Randy -- standards used to be about users trying to get vendors to
         interop so that we have choices.  IETF is now about vendors
         trying to sell us stuff.  I didn't come out of today better
         at protocol design or with better knowledge about what to do
         for users.
         

- - Next steps (5 min)
- - AOB