Re: [arch-d] [IAB] Call for Comment: <draft-iab-for-the-users-02> (The Internet is for End Users)

[Joel Halpern Direct <[e-mail]>]

Yes, put simply, I think that declaring that the user's interest takes 
precedence over everything else is wrong.  You or the IAB as a whole are 
of course free to come to different conclusions, but I think it is 
actually very helpful to view it as a tussle with a veriety of itnerests 
that need to be respected.

If the IAB really wants to say that they think the tussle is over, I 
would expect to see the names of approving IAB members at the end of the 
document when this is done.  If, on the other hand the IAB considers 
this a useful perspective but not necessarily one they all agree with, 
then I would expect that to be clearly stated.

Personally, if the IAB takes the stance that this is a done deal, and 
the tussle is over, that will decrease the likelihood of my bothering to 
pay attention to the specifics (to the extent they exist) in the 
document.  Your choice.

Yours,
Joel

On 2/6/2020 9:28 AM, Martin Thomson wrote:
> Hi Joel,
> 
> On Thu, Feb 6, 2020, at 01:24, Joel M. Halpern wrote:
>> But the underlying tone that the End User concerns should take
>> precedence over everything else seems to be incorrect.
> 
> Are you suggesting that it is incorrect in that it is not a provable statement in a mathematical sense and more a value judgment?
> 
>  From my perspective, that value judgment is right (in the moral sense) and that we should be saying this.  Having the IAB say this is also important, though I would of course be happier if the IETF said the same.  This statement is consistent with similar statements from the IETF (as referenced, RFC 3935), and other organizations (see for example the W3C position <https://www.w3.org/TR/html-design-principles/#priority-of-constituencies>).
> 
> As a letter to the IETF, I hope that it has weight, but it does not set out to have any sort of normative force.  Choosing a different basis from which to design a protocol would not automatically disqualify it from consideration for publication on the IETF stream, though I hope that people would be motivated to ask pointed questions or contest something that, for instance, might favour operations to the detriment of individual interests.  Arguably, we already do that; see RFCs 7871 and 7974 for published cases of where that happened with different results.
> 
> You cite some examples that seem to imply that you think that this general assertion is wrong in general, namely connected "stuff" and people who operate the network (though not those who build and design it, curiously).  Both of those concerns are addressed in the draft.  To sharpen those points: network operations exist to serve the network, which exists to serve the users of that network; devices exist to serve a range of use cases, which generally ultimately have users also.  Draw a longer chain and the value to people becomes more diffuse and abstract; see also <https://en.wikipedia.org/wiki/Trickle-down_economics>.  But the draft also recognizes the potential for instances where the balance of interests might end up favouring other entities.
> 
> I do see this as a partial refutation of the tussle theory, to Bernard's point, but I don't see a need to address that more directly.
> 
> It's true that we cannot always predict the effect of our choices, but it seems to me that there is a prevailing attitude that says we MUST NOT consider things like social implications.  That to me is where we go wrong.  Good engineering considers all the requirements, not just the convenient ones.
> 
> Cheers,
> Martin
>