ietf
[Top] [All Lists]

Re: Gather Profiles/Resumes [was Re: call for ideas: tail-heavy IETF process]

2013-05-16 10:47:15
From: Thomas Narten <narten(_at_)us(_dot_)ibm(_dot_)com>

What I do think the IETF should do is *require* that participants
identify themselves. That means knowing who they are (a name and email
contact) and an affiliation. For 80% of the participants, this info is
not very hard to figure out (see below). But we also have participants
that use obscure email handles that don't correlate to anything
obvious, whether a real person or to a name in the list of registered
attendees, etc. I suspect these folk are *not* intenending to be
anonymous participants, but in practice they are.

And yes, knowing who someone is, their background and who they work
for is important to me in figuring out how to guage their input. E.g.,
I would likely pay more attention to an operator's comments on a
proposed use case than from someone else.

I believe that this is a less good idea than first appears.

One objection is that the concept of "affiliation" is less simple than
it appears.  People normally expect it to mean "Who is your employer?"

As John mentions, it can mean "What is the organization whose
interests you are promoting?"  But in regard to your observation about
use cases, it can also mean "What area of technology do you have
experience in?"

In many cases, "affiliation" is related to "Who is paying for your
attendance?"  But there are situations where your employer may be
willing to pay for your attendance but doesn't want you to be seen as
officially representing them.

And (at least in common-law countries) one can simply invent an
(unincorporated!) company name and claim it as your affiliation.

In practice, the way we deal with these problems is that we consider
everyone to be individuals, rather than as "the representative of
company XYZ", and it takes time to build a reputation.  One is thus
unwilling to sacrifice that expensive reputation to advocate a poor
solution because one's employer favors it.

    3) Google names, look at authorship info in RFCs, linked in,
    etc. Works in a lot of cases, but is sometimes more work than
    seems appropriate.

If you can't find any information about them, they have no reputation.

Dale

<Prev in Thread] Current Thread [Next in Thread>