Re: Making Decisions in the IETF

Watson Ladd <[email protected]> Wed, 29 Jul 2026 11:00:28 -0700
Newsgroups gmane.ietf.general
Message-ID <CACsn0cm9=qNb=RgUnuvWwWYdokXWXb2_XSMvMeji-UUkgWpKOw@mail.gmail.com>
--000000000000bcce7d0657c3bb0b
Content-Type: text/plain; charset="UTF-8"
Content-Transfer-Encoding: quoted-printable

On Wed, Jul 22, 2026, 4:38=E2=80=AFPM Benjamin Kaduk <[email protected]> wrote:

> Hi SM,
>
> On Wed, Jul 22, 2026 at 04:04:00PM -0700, S Moonesamy wrote:
> > Hi Mark, Pete,
> >
> > I read the draft about "Making Decisions in the IETF".
>
> I assume this refers to draft-nottnick-ietf-decisions?
>
> > Section 1 of the draft states that the guidance in RFC 2418 has served
> the
> > IETF well for more than thirty years.  It then goes on to argue that th=
e
> > IETF has grown and the decisions made are more than ever relevant to
> > society.
> >
> >   (a) It's not often that one finds guidance which has served a group
> well
> > for over
> >       thirty years.  Why is the guidance no longer working as it did
> thirty
> > years ago?
>
> I'm not either Mark or Pete, but I would argue that we have at least one
> worked example of the original guidance not working well
> (draft-ietf-tls-mlkem) and it seems totally reasonable to use that as a
> prompt to provide better guidance.
>

In my view the lesson learned is "don't": avoid situations where one has to
pick one of two options unless absolutely necessary. This especially
applies when there are few technical criteria to apply.



> >   (b) What changed in the group as it grew?
> >
> >   (c) What is the difference in the decisions made thirty years ago and
> > those taken
> >       nowadays?
> >
> > Section 1.1 has the following: "decision making by voting is
> inappropriate
> > -- it would make our processes vulnerable to rule by majority and vote
> > stuffing."  I followed a few working groups today.  There was voting in=
 a
> > few of those working groups.  If I read the text from the draft
> literally, I
> > would conclude that there was something inappropriate going on.
>
> I don't think your description of what you observed provides enough
> information to be able to really answer the question.  It might indicate
> something inappropriate, or it might just be a proxy for soliciting
> objections and assesing support, and such a proxy is probably fine in the
> absence of objections.
>
> > Section 2 of the draft has the following: "groups SHOULD maintain a
> record
> > of decisions".   Are there cases where a group does not have to record =
a
> > decision?
> >
> > Section 3 of the draft states that adoption does not consensus.  I
> thought
> > that little gem was fairly well-known.  That is not reflected in the
> various
> > mailing list discussions which I have seen.
>
> Are you proposing a change to the draft?
>
>
> Having been prompted to look at the draft, I also have some comments not
> inspired by SM's review.
>
> On the whole, I think it's broadly aligned with my understanding of what =
a
> WG chair is supposed to be doing when gauging consensus.
>
> However, I find parts of 2.3 (Handling Objections) underwhelming, largely
> relating to the assumption that any given objection must either be upheld
> or discounted, and also about what constitutes grounds for discounting an
> objection.
>
> In particular, the document is silent on the class of scenarios where the=
re
> is a tradeoff between two or more options that have different strengths a=
nd
> weaknesses, with objections raised as to the weaknesses of each option.  =
I
> do not think it is correct (in the general case) to say that all the
> objections are upheld and the WG can do nothing; the WG should be able to
> make some decision.  If we take that as the premise, that leads to a
> situation where we have a valid objection that we choose to ignore for
> other reasons, e.g., as part of a tradeoff between options.
>
> When consensus callers are faced with such a tradeoff situation, I think
> the guidance should be roughly that the tradeoff needs to be considered i=
n
> the scope of the identified use cases (which implicitly weakens the
> position of unidentified use cases), probably weighted by the number of
> impacted end users if needed (see RFC 8890 for the IAB's take), while
> trying to enable as many use cases as possible to achieve some kind of
> useful (even if not optimal) result and minimize the number of use cases
> that experience drastic harm from the given choice (as compared to some
> other alternative choice).  I would also generally expect the tradeoff
> to be described in some form in the resulting document (presumably with
> some exceptions possible).
>
> I also think it would be helpful to give guidance to consensus callers on
> how to convey the results of the decision to the group, e.g., what
> information needs to be included, how to describe the decision process,
> what tradeoffs and use-cases were weighed against each other.  IIRC it to=
ok
> a significnat amount of time to produce summaries like
> https://mailarchive.ietf.org/arch/msg/tls/2QHMMX9g5vpFJaeH0okj40N3WjQ/ bu=
t
> I only remember ever getting positive feedback about having produced them=
.
>
> -Ben
>
>

--000000000000bcce7d0657c3bb0b
Content-Type: text/html; charset="UTF-8"
Content-Transfer-Encoding: quoted-printable

<div dir=3D"ltr"><div dir=3D"auto"><div dir=3D"auto"><br></div><div class=
=3D"gmail_quote" dir=3D"auto"><div dir=3D"ltr" class=3D"gmail_attr">On Wed,=
 Jul 22, 2026, 4:38=E2=80=AFPM Benjamin Kaduk &lt;<a href=3D"mailto:kaduk@m=
it.edu" target=3D"_blank">[email protected]</a>&gt; wrote:<br></div><blockquote=
 class=3D"gmail_quote" style=3D"margin:0px 0px 0px 0.8ex;border-left:1px so=
lid rgb(204,204,204);padding-left:1ex">Hi SM,<br>
<br>
On Wed, Jul 22, 2026 at 04:04:00PM -0700, S Moonesamy wrote:<br>
&gt; Hi Mark, Pete,<br>
&gt; <br>
&gt; I read the draft about &quot;Making Decisions in the IETF&quot;.<br>
<br>
I assume this refers to draft-nottnick-ietf-decisions?<br>
<br>
&gt; Section 1 of the draft states that the guidance in RFC 2418 has served=
 the<br>
&gt; IETF well for more than thirty years.=C2=A0 It then goes on to argue t=
hat the<br>
&gt; IETF has grown and the decisions made are more than ever relevant to<b=
r>
&gt; society.<br>
&gt; <br>
&gt;=C2=A0 =C2=A0(a) It&#39;s not often that one finds guidance which has s=
erved a group well<br>
&gt; for over<br>
&gt;=C2=A0 =C2=A0 =C2=A0 =C2=A0thirty years.=C2=A0 Why is the guidance no l=
onger working as it did thirty<br>
&gt; years ago?<br>
<br>
I&#39;m not either Mark or Pete, but I would argue that we have at least on=
e<br>
worked example of the original guidance not working well<br>
(draft-ietf-tls-mlkem) and it seems totally reasonable to use that as a<br>
prompt to provide better guidance.<br></blockquote></div><div dir=3D"auto">=
<br></div><div dir=3D"auto">In my view the lesson learned is &quot;don&#39;=
t&quot;: avoid situations where one has to pick one of two options unless a=
bsolutely necessary. This especially applies when there are few technical c=
riteria to apply.</div><div dir=3D"auto"><br></div><div dir=3D"auto"><br></=
div><div class=3D"gmail_quote" dir=3D"auto"><blockquote class=3D"gmail_quot=
e" style=3D"margin:0px 0px 0px 0.8ex;border-left:1px solid rgb(204,204,204)=
;padding-left:1ex">
<br>
&gt;=C2=A0 =C2=A0(b) What changed in the group as it grew?<br>
&gt; <br>
&gt;=C2=A0 =C2=A0(c) What is the difference in the decisions made thirty ye=
ars ago and<br>
&gt; those taken<br>
&gt;=C2=A0 =C2=A0 =C2=A0 =C2=A0nowadays?<br>
&gt; <br>
&gt; Section 1.1 has the following: &quot;decision making by voting is inap=
propriate<br>
&gt; -- it would make our processes vulnerable to rule by majority and vote=
<br>
&gt; stuffing.&quot;=C2=A0 I followed a few working groups today.=C2=A0 The=
re was voting in a<br>
&gt; few of those working groups.=C2=A0 If I read the text from the draft l=
iterally, I<br>
&gt; would conclude that there was something inappropriate going on.<br>
<br>
I don&#39;t think your description of what you observed provides enough<br>
information to be able to really answer the question.=C2=A0 It might indica=
te<br>
something inappropriate, or it might just be a proxy for soliciting<br>
objections and assesing support, and such a proxy is probably fine in the<b=
r>
absence of objections.<br>
<br>
&gt; Section 2 of the draft has the following: &quot;groups SHOULD maintain=
 a record<br>
&gt; of decisions&quot;.=C2=A0 =C2=A0Are there cases where a group does not=
 have to record a<br>
&gt; decision?<br>
&gt; <br>
&gt; Section 3 of the draft states that adoption does not consensus.=C2=A0 =
I thought<br>
&gt; that little gem was fairly well-known.=C2=A0 That is not reflected in =
the various<br>
&gt; mailing list discussions which I have seen.<br>
<br>
Are you proposing a change to the draft?<br>
<br>
<br>
Having been prompted to look at the draft, I also have some comments not<br=
>
inspired by SM&#39;s review.<br>
<br>
On the whole, I think it&#39;s broadly aligned with my understanding of wha=
t a<br>
WG chair is supposed to be doing when gauging consensus.<br>
<br>
However, I find parts of 2.3 (Handling Objections) underwhelming, largely<b=
r>
relating to the assumption that any given objection must either be upheld<b=
r>
or discounted, and also about what constitutes grounds for discounting an<b=
r>
objection.<br>
<br>
In particular, the document is silent on the class of scenarios where there=
<br>
is a tradeoff between two or more options that have different strengths and=
<br>
weaknesses, with objections raised as to the weaknesses of each option.=C2=
=A0 I<br>
do not think it is correct (in the general case) to say that all the<br>
objections are upheld and the WG can do nothing; the WG should be able to<b=
r>
make some decision.=C2=A0 If we take that as the premise, that leads to a<b=
r>
situation where we have a valid objection that we choose to ignore for<br>
other reasons, e.g., as part of a tradeoff between options.<br>
<br>
When consensus callers are faced with such a tradeoff situation, I think<br=
>
the guidance should be roughly that the tradeoff needs to be considered in<=
br>
the scope of the identified use cases (which implicitly weakens the<br>
position of unidentified use cases), probably weighted by the number of<br>
impacted end users if needed (see RFC 8890 for the IAB&#39;s take), while<b=
r>
trying to enable as many use cases as possible to achieve some kind of<br>
useful (even if not optimal) result and minimize the number of use cases<br=
>
that experience drastic harm from the given choice (as compared to some<br>
other alternative choice).=C2=A0 I would also generally expect the tradeoff=
<br>
to be described in some form in the resulting document (presumably with<br>
some exceptions possible).<br>
<br>
I also think it would be helpful to give guidance to consensus callers on<b=
r>
how to convey the results of the decision to the group, e.g., what<br>
information needs to be included, how to describe the decision process,<br>
what tradeoffs and use-cases were weighed against each other.=C2=A0 IIRC it=
 took<br>
a significnat amount of time to produce summaries like<br>
<a href=3D"https://mailarchive.ietf.org/arch/msg/tls/2QHMMX9g5vpFJaeH0okj40=
N3WjQ/" rel=3D"noreferrer noreferrer" target=3D"_blank">https://mailarchive=
.ietf.org/arch/msg/tls/2QHMMX9g5vpFJaeH0okj40N3WjQ/</a> but<br>
I only remember ever getting positive feedback about having produced them.<=
br>
<br>
-Ben<br>
<br>
</blockquote></div></div>
</div>

--000000000000bcce7d0657c3bb0b--