Re: [hrpc] I-D Action: draft-irtf-hrpc-political-05.txt

[Niels ten Oever <[e-mail]>]

Thanks for this detailed review Stephen.

Because there was some quiet on the list (which might indicate some level of agreement (either with -05 and/or Stephen's comments, but I'll leave that to others to decide), I wrote a detailed response underneath and made changes in version -06.

On 9/19/19 10:46 PM, Stephen Farrell wrote:
> 
> Hiya,
> 
> I had a read of -05. I didn't read (but just skimmed) all
> the other comments before doing that so apologies if my
> comments repeat (or utterly contradict:-) others that've
> been sent to the list already.
> 
> Cheers,
> S.
> 
> General:
> 
> - Overall, I don't think this is ready. I do think it could be with
>   some more work, maybe another rev or two. Or it might take longer,
>   not sure.
> 

Working on next rev :)

> - If the authors didn't want to try aim for RG consensus on all of
>   the text, then I think that'd make it easier to get an RFC
>   published. I'd be ok with that, though I still think it'd need
>   another rev or two (but the changes might be easier/less).
> 

I'll leave that decision to the RG chairs. I am happy with either. But I do see this as a product of the discussions in hrpc, so publishing it outside of that would not do justice to the document and the people that contributed to it imho.

> - I'd argue to entirely re-do or delete section 5. I see why it's
>   included but I think in the end it didn't work out to be useful in
>   the same document as section 4.
> 

I removed section 5, but kept two para's and added them to the intro.

> Detailed comments:
> 
> - "undisputed" in the abstract is likely to be disputed:-) By
>   readers, I mean. s/undisputed/seems evident to the authors/ or
> something might be better.
> 

I do think it _is_ undisputed that the standards creation process is political and that they can be used for political means, right? But to make it go down easier:

s/undisputed/generally agreed 

> - intro: the Callon quote could do with a reference so readers can go
>   find more context if they want. The first sentence of that quote
> could mean lots of different things and I've no idea what exactly was
> meant - I'd ditch that sentence or maybe the entire quote.
> 

Callon is one of the parents of Science and Technology Studies, but I do agree that outside of the discipline, this quote might seem open for a lot of different interpretations. I removed it, with pain in my heart, and replaced it with two more specific quotes from the literature on standards.

> - intro: "affordances" - I've heard that term a number of times, but
>   I don't think I really understand it. Also - how are users
> "shaped"? I think you could have said that "people are affected by"

Suggestion accepted and changed.

> the technology, left it at that and been clearer.  (I see you define
> that in section 2, so I do get it now that I've gotten that far.
> Still might be better to avoid the forward reference if possible.)
>> - intro: does the document really "seek to answer" the question or
>   rather try to explore the question? The latter seems more
> tractable.
>

Accepted and changed.

 
> - intro: "The design of the Internet through protocols and
>   standards..." That seems to ignore the "running code" aspect which
> is critical to understanding how the Internet has arisen and evolves.
> 

It doesn't say 'exclusively through', but I appreciate the point. I changed it into:

The design of the Internet, and its codification through protocols and standards, is a technical issue with great political and economic impacts, as is described in {{RFC0613}} and {{RFC3271}}. 

> - The URL for [BramamI] doesn't work for me - my browser times out on
>   that. 

Fixed BramanI and Braman two, unfortunately Texas A&M does not offer TLS for this URI

(And why http and not https? :-) 


> As below, it'd be better to
> say a bit more about i18n here - it's fairly obvious but no harm
> expanding the bullet beyond one word.
> 

Am afraid this would deviate from the main point, since these are just illustrations. Previously the bullets were comma's, but Ted and Paul objected against that because with the references it impacted the readability.

But it indeed seems weird to have one word bullets, so I added one sentence summaries.

> - intro: The reference to RFC101 (better that than RFC0101 btw:-) is
>   unclear to me. There are many occurrences of the word access in
> that RFC so what's meant by the bullet point here? It's not clear to
> me.
> 

'how people are able to access the network, and who has control {{RFC0101}};'

> - intro: same point about rfc49 - that has two mentions of "security"
>   neither of which seem that relevant to this draft, so maybe say
> what/why that reference is for. (And there may be a better one, if
> what you've done is try find the earliest reference, that might not
> be the best thing.)


Added better reference and added privacy, as well as a description:

* privacy and security {{BramanIII}}; 

> 
> - intro: same points for the other early RFC references and I guess
>   [BramanII].
> 
> - intro: "This has been clearly shown in the work done by Sandra
>   Braman..." - what "this" do you mean has been shown? It's not clear
> to this reader.
> 

Changed into: 

Sandra Braman has foregrounded these political consideration in historical RFC in her extensively analysis of these documents {{BramanII}}.


> - section 3: Suggest s/Answering/Exploring/
> 

Accepted and changed

> - section 4: RFC1087 says "The Internet is a national facility..." so
>   I wonder if it's really so obvious a basis for the argument here.
> The problem is that 1087 kinda says "be ethical or we'll tell the
> Feds" and there aren't really equivalent Feds for today's Internet,
> so the sentence here could be a bit misleading.  Maybe better to
> qualify the reference here with something like "... even if the
> Internet in 1989 was largely a US thing, similar problems and
> requirements for ethical behaviour ought still apply." Not sure.
> 

Hmmm. This sentence does not say that the Internet is still the same, right? Just that is has been extensively discussed in earlier days?

But I've changed the sentence into:

Ethics and the Internet was already a topic of an RFC by the IAB in 1989 {{RFC1087}}, when the Internet was still looking entirely different.

> - 4.1: I note the references here are from 1991 and 1992 - isn't
>   there a more recent expression of this position that could be
> quoted? It left me wondering if the referenced authors might have
> changed their positions since (I've no idea if they have or not, just
> wondering, as my own ideas now are not what they were then:-)
> 

This is a relatively classical expression of this position, that is why I went with the classical definition of it (and the author expressing it). 

> - 4.2: This one is closest to my own position, but I wouldn't go
>   along with "requires" which is far too strong - for me it's more
> like this position "implies that protocols could be evaluated for..."
> 

Changed.


> - 4.3: Even if the Internet is too complex, that doesn't imply
>   anything for any one protocol. 

It means that the interrelation between protocols in hard to predict.

The rest of the 4.3 text also seems
> unclear to me. I'd say a rewrite of this section would be good.
> 

Added text and sought to increase readability.

> - 4.4: "The network..." do you mean the Internet or one of the
>   networks that make up the Internet here? Or do you mean any network
> of networks? I assume you mean any network of networks, like the
> Internet but I'm not sure.
> 

Fixed that by making it all 'network of networks'

> - 4.4: This seems to finish early or something. The end of the
>   section talks about social and technical but never says political
> so I'm unsure what a proponent of this position might believe vs. the
> other positions in section 4.
> 
> - 4.5 (1): I've no clue what the emotional bias of a medium such as
>   audio might be, or am I missing the point?
> 

What is meant here that a newspaper instills a very different perception than radio does. But to decrease misunderstanding I removed the second part of that sentence.

> - 4.5 (2): "speed of their information" is an odd phrase, and I'm not
>   sure what's meant.
> 

I think we can safely assume high bandwith and low latency is what is meant here.

> - 4.5: IMO the raven process was an attempt to be technical and
>   apolitical, at least as understood by some participants.  It's
> likely incorrect to give the impression that everybody thinks that
> the raven process was "largely political." (I do agree that many
> people think that.) 
> Similarly, while I agree that what "engendered"
> 7258 was Snowdonia, I do not agree that 7258 requires any
> justification that is non-technical (i.e. IMO 2804 and 7258 and 1984
> can all be sufficiently justified on purely technical grounds).
> Again, I think the change to make here is to say that someone
> espousing the position of 4.5 might agree with the current text, but
> that those statements about 2804 and 7258 are not universally
> accepted.

OK, I propose the following:

The Raven process in which the IETF refused to standardize wiretapping -- which resulted in {{RFC2804}} -- was an instance where an international governance body took a position that was perceived by many as political, although driven by a technical argument. 

> 
> - 4.5: Neither 2804 nor 7258 are protocol specifications.  For me
>   that significantly weakens the overall argument proposed in 4.5.
> Surely soeomne who espouses this position could provide at least one
> actual protocol example?  Presumably something that refers to 2804 or
> 7258 might suffice, but without that, there's a hole in the argument.
> (Not that I agree with the argument in 4.5, but I feel better
> disagreeing with something that isn't holed beneath the water line:-)
> 

Added:

The Raven process in which the IETF refused to standardize wiretapping -- which resulted in {{RFC2804}} -- was an instance where an international governance body took a position that was perceived by many as political, although driven by a technical argument. The process that led to {{RFC7258}} is similar: the Snowden disclosures, which occurred in the political space, engendered the IETF to act. While {{RFC2804}} was a statement about how a protocol for wiretapping would _not_ be developed, {{RFC7258}} was a statement that contributed to the development of protocols such as {{RFC7858}}, {{RFC8226}}, and {{RFC8404}}. The impact of political tensions on protocol development is summarized in {{Abbate}} who says: "protocols are politics by other means," emphasizing the interests that are at play in the process of designing standards.


> - 4.5: "consumer class" seems wrong to me. I don't believe there's
>   what I'd call a "consumer class" for NTP for example. I'm not clear
> from the phraseology as to whether the authors are asserting that
> espousing 4.5 means agreeing with Winner here or not. In any case, I
> think Winner is wrong here, for some protocols.

Removed 'consumer class', to avoid discussions about class, groups, and consumerism. I would offer the following:

Those who had the power to introduce a new technology also had the power to largely frame the uses of the technology "with new practices, relationships, and identities supplanting the old, --- and those who had the wherewithal to implement new technologies often molded society to match the needs of emerging technologies and organizations." {{Winner}}.

> 
> - section 5: "the different existing positions" - that's a claim that
>   secction 4 has the full set. My bet is that I could easily find an
> IETFer who espouses none of the above. s/the/some/ is better. Or you
> could say those are the postions that were considered in this work.
> 

Removed

> - section 5: s/were created/are created/ and "formal" isn't quite
>   right - TLS1.3 is getting closer to formal since bits of it were
> formally verified but it is not formally specified in one well-known
> sense of the term. s/formal// would be my suggestion. (I guess you
> just mean "written down" by this use of "formal" which seems
> unnecessary to me.)
> 

Removed

> - section 5: "100% royalty-free"? What? I'm misundertanding something
>   here. We had years of RSA and ECC being encumbered to the detriment
> of the Internet IMO, but 100% seems just wrong or else must be a
> badly stated statement of something else?
> 

Removed

> - 5.1: "enhance competition" may be true sometimes but is a) not
>   always true (I'll try assert NTP again as an example) and b) is far
> down my list of why to do work in the IETF, but is presented as if we
> all think this is the most important thing ever.
> 

Removed

> - 5.1: the discussion of "de facto" seems broken to me. Consider RFC
>   3986 vs. whatwg for a real example that doesn't seem to fit your
> description. (And I barf again at the mention of competition law for
> such things as URL syntax:-)
> 
Removed

> - 5.1: mention of ISO, BIS... etc with no references or URLs seems
>   wrong. (I've no idea what ABNT is for example, but I guess it's not
> what first shows up in my search [1].)
> 
>    [1] https://acronyms.thefreedictionary.com/ABNT

Removed

> 
> - 5.4: Are the authors agreeing with [Nadvi] or not? I think you
>   ought be clear. I'm not sure I'd agree that applies to the IETF
> though, as IETF participants don't, in general, seem to me to be
> well-decscribed as "transnational stakeholders" even though some or
> many of them may be. I don't feel like one of those anyway;-)

I've added that these are not exclusively transnational corporations.

> 
> - 5.4: The big [RogersEden] quote goes against IETF lore. I forget
>   where we have that written down in an RFC but I think we do. (Where
> "that" is "don't just end up with the union of the input proposals
> but instead try find the right outcome" or something like that.

Removed


> 
> - Section 5 generally: I was left thinking "huh, what was the point
>   of that section" when I finished reading that. I think you could
> mostly delete the entire section and be ok.
> 

Done

> - Section 6: I'd be happier if statements such as your opener did not
>   depend on buying books. Haven't Russel or Abate ever written these
> arguments down in something openly accessible?  

Nopes, they did not publish their papers in open access journals. I do have a problem with that, but I am afraid that is beyond my power to fix. 

> (And BramanII is
> still a broken link.) 

Fixed

> I don't think you've made the case via text or
> accessible references that this opening conclusion statement is
> correct. (I may agree with it, but I don't think you can justify it
> based on this.)
> 
> - Section 6: "undisputed" again - I dispute that that's the right
>   term.

s/undisputed/generally agreed
> 
> - Section 6: I do not accept the [Russell] quote as being correct but
>   haven't bought his/her book. I don't think you ought present that
> quote as being an accepted statement.

Made that more clear.

> 
> - Section 6: I dispute that the implementation of protocols "is"
>   political. I wrote some code last night for an I-D. That was not a
> politcal act (I assert).
> 

Would adding 'often' help?

While understanding that 'standards emerge from contested contexts, they immediately function as a means of control within the political and economic order' {{Russell}}, protocols and standards as abstract isolated artefacts might not be political, but their design, development, deployment, and implementation often is.


> - Section 6: I dispute that "protocols can never be understood"
>   without some inherently political framing. (If that's what you
> meant.) I can teach kids about TLS and they can understand the
> protocol without all the political add-ons. As it happens I do think
> they better understand what TLS is and how it's used, if I include
> some bits of that, but I don't believe it's needed for them to
> understand the protocol itself.
> 

I propose:

'Therefore we might need to give a qualified answer to the research question, in the sense that protocols can only be understood in part outside of their actual shaping, use, and applied function, which is political.'


> - Section 6: I dunno if I agree about the future research.

Would you care to elaborate? Happy to sketch other paths forward. 

Thanks again and happy to discuss.

Best,

Niels

PS If other people think their point(s) is a/are showstopper(s) and should be addressed, feel free to re-up your comments or message me off-list so I can address them in a following version.


> 
> 
> _______________________________________________
> hrpc mailing list
> [e-mail]
> https://www.irtf.org/mailman/listinfo/hrpc
> 

-- 
Niels ten Oever
Researcher and PhD Candidate
DATACTIVE Research Group
University of Amsterdam

PGP fingerprint	   2458 0B70 5C4A FD8A 9488  
                   643A 0ED8 3F3A 468A C8B3