Re: [arch-d] Draft IAB conflict of interest policy

[John C Klensin <[e-mail]>]

--On Thursday, January 16, 2020 16:10 +1300 Brian E Carpenter
<[e-mail]> wrote:

>>     "Can the work of the IAB that you contribute to impact
>>     future financial results of your employer / sponsor ?"
> 
> I think that's easy. Read the arguments that you, as a
> candidate, made to your employer to justify you spending time
> and travel budget on IAB membership. If those arguments do not
> explain how IAB membership will positively affect your
> employer, answer "No". But in that case, I rather doubt that
> your employer would have allowed you to accept the nomination.
> Mine certainly wouldn't have.
> 
> In other words, the only credible answer to this question is
> "Yes".  
>>     If one would answer YES, does that constitute a COI ?
> 
> Poetntially, yes. Even a self-employed engineer has this
> potential COI. It's something I lived with as a WG chair, IAB
> member, IESG member, ISOC Board member, and for that matter in
> completely other non-profit roles at various times.  

Brian,

I think you are right, but I also have to agree with Randy.
Someone working for an organization that is willing to have them
participate in the IETF because of a commitment to the general
health of the Internet and that is willing to accept their
acting against the best short-term interests of the company's
products is a rare situation, but one that several of us have
been in at one time or another.  In some ways, that is the
ideal, at least as long as we don't increase costs of
participation enough that only the largest organizations can
play.  But, even then, if the individual is nearly as
opportunistic as some of the comments in this thread imply that
a policy needs to look out for, then the individual might still
reason that, if the company's products do well this year, then
there will be bigger bonuses next year and therefore be biased
toward those products. 

I stopped commenting several days ago because I tried to write
an explanation of why someone with projects or clients that
could not be made public might still be an appropriate
participant in the IAB.  I realized that the explanation is too
complicated if the policies have to be simple.  I also realized
that, if I were working for a company whose primary business was
the making of soap or breakfast cereal, it might be hard to
identify conflicts of interest without an in-depth job and
departmental description and companies often consider that sort
of information (and organization charts beyond some level of
detail) to be sensitive ... a situation not very different from
a consulting contract that barred either party from disclosing
the relationship without mutual agreement (and that might have
been insisted on by the consultant rather than the company).
FWIW, I have never had any trouble getting permission to
disclose arrangements that are, formally, non-public or
confidential to Nomcoms: but the rules and the culture there
have seemed offer more than adequate protection of the
information.

Where this all leads me is back to a point that others have made
in different ways.  For the technical work of the IAB, we want
and need as much diversity of opinions, technical expertise, and
perspective as we can get.  That is, as long as the people
involved really are experts and their expertise includes knowing
what they don't know.  IAB members might be biased as much or
more because of personal technical preferences than because of
financial interests (however described or measured).  If we
decide the former are ok and the latter aren't, we are not only
going to find ourselves splitting hairs and exploring rat holes,
but do each other --and the IAB's historical core technical
mission-- a disservice.  

As an extreme, and I hope silly, example, if the Nomcom, in its
wisdom, decided to appoint someone to the IAB who was the main
advocate of a protocol or device that claimed to make physical
limits on IP based on the speed or protons or electrons in a
vacuum irrelevant, in technical discussions, I'd want the IAB to
hear and evaluate that perspective and to do so independent of
whether the individual stood to make zillions of dollars if the
protocol proved successful.  I'd also want the IAB to hear from
the members best equipped to explain why, for Internet purposes
at least, protocols or devices that require trans-light speeds
to work properly are not going anywhere, independent of what
they might stand to gain from upholding those positions.  If
either or both of those types of positions should not be
represented in IAB discussions, then we are a problem with the
way nomcoms work or with that absence of an effective recall
mechanism, neither of which will be helped by a complex and
hair-splitting COI policy.

If we need very different policies for the IAB's technical and
architectural work than for its more administrative or
representational functions, it is likely that there are
different skill sets for those functions as well.  And maybe
someone who has real or imagined conflicts should be isolated
from the latter even while being a valuable asset for the
technical and architectural work.  If that is the case, maybe it
is time to reconsider the IAB and whether some of its functions
belong somewhere else.  The perceived need for a formalized COI
policy may be another argument that it is time to rethink the
IAB, but it is not a solution.

I think the bottom line is that this policy, at least as
currently written, is unlikely to be helpful in difficult cases
and may easily be damaging.  Nomcoms  need to select IAB members
who hold openness and transparency in sufficient respect, and
who have a strong enough of personal ethics, that issues that
might affect their judgment on technical issues will be
disclosed to other IAB members and, if appropriate, the
community and who are willing to keep in confidence anything
that is told them in confidence.  If Nomcoms can't be counted on
to do that, probably nothing else matters because people with
less integrity (a problem I don't think we have seen) can always
lie about their conflict.  If Nomcoms cannot do that reliably
with the tools they have available, then we have a different
kind of problem, but, again, complex COI policy statements are
not going to fix it.

And, again, if we have given the IAB responsibilities
--financial, administrative, or otherwise and direct or
indirect-- that do require this sort of statement, let's work at
separating those responsibilities as incompatible with what I
still believe is the IAB's primary role, rather than things it
ended up with post-POISED because we didn't have a better plan
and other non-architectural tasks that have accumulated since.

regards,
   john