At IETF-96, we had a discussion about both the chosen starting time but also
other issues in the way that we set up the schedule, and provide time for
either WG or other discussions.
This email provides some information about the options the IESG has in setting
the schedule and related tradeoffs.
In general, barring any specific venue limitations, the IESG can independently
adjust a number of different variables in the schedule, while assuming the same
overall meeting week length. We can adjust the length of the day and overall
number of hours in the schedule. The IESG sets only the formal meeting
schedules, so the tradeoff here is what fraction of the days are spent in
formal meetings vs. unscheduled time that might be used for design team
meetings and other more unstructured collaboration.
We can also adjust the specific timing, subject to local food availability and
other conditions. And we can adjust the parallelism, i.e., number of tracks.
This will impact conflicts but also our capacity to host meetings that the WGs
have requested. Finally, we can adjust the number of plenaries, with the recent
trend being holding exactly one plenary during the week.
Overall, our perception at the IESG is the following:
o The start time provokes some opinions. For IETF-97 we plan to set this so
that lunch and dinner times in the local situation are convenient. It might be
useful to poll the community about this as well, but we'd like to suggest that
it may be more important to focus on the bigger issues such as
unstructured/structured time. At this time, what we know is that we will have
plenty of restaurants around us, that there may be some lines at 11:30, some
restaurants close at 21:30 while others stay open 24 hours. We are still
deciding what to do about the start time, though one possibility is to start at
9:30. A lot depends also on how many sessions WGs request, or how many official
total meeting hours we want to have.
o Knowing the ending time on Friday is important for your flight arrangements.
We do plan to start early on Friday, have three hours of sessions and one small
room changeover break. As a result Friday schedule will end at 12:20.
o We believe the community has told us to not increase the plenary time. The
existing plenary time should be used primarily for technical topic (when
available) or community discussion. There will be some amount of program in all
of our plenaries, but the substantial reports have largely been moved to online
consumption.
o We also believe the community has told that arranging more unstructured time
for work to happen would be useful. Is this a correct conclusion? The exact
forms of how we might go about this are subject to furher discussion, see
below. One of the big question marks here is the effect of any unstructured
time to remote or English-as-foreign-language participants.
o We have also heard repeated calls for stricter time allocation control by
the IESG, in that only some working groups or maybe no working groups should be
allowed extensive use of meeting hours (2x3hrs etc). We're not quite sure where
the community is on this though, and it is also important that we get to do
work when we do meet face to face. Our current belief is that it is better to
deal with this question as part of the overall time allocation question, i.e.,
how much time gets allocated. The IESG is in charge of ensuring that within the
time reserved for WGs, the allocation is fair in the sense of giving the time
where it is needed and demonstrably useful. Including, for instance, the WG's
track record from the time since the previous IETF meeting.
With regards to the possible increase of unstructured time in the meetings, we
would like to do some experimentation, as we do not yet have experience of how
time might get used.
The possible experiments that we've identified are:
1. The total number of hours available for WG meetings during the week. We
could either keep this capped at what we had in Berlin. Alternatively, we can
reduce the overall number of hours so that we can have, for instance, more
design team meetings. We can start later or finish earlier one day, as an
example.
2. Encouraging working groups try unstructured meeting time in one of their
slots. This time could be used for free-form discussion in the whole group, an
“unconference" style idea session in the different corners of the room, or a
hacking session devoted to the WG's technology. All up to the WG chairs to
decide. We’ve reached out to the working group chairs for their ideas in this
space.
Note: Seoul we will have tighter than usual space constraint. We can
unfortunately not devote additional rooms, so anything we do needs to run under
the usual room allocations. However, we will have more space in Chicago, so
some bigger experiments or changes could be considered there.
We would like to solicit some feedback:
Q1. Please confirm that the community wishes that we arrange more unstructured
time for work to happen.
Q2. Would you like to either keep the amount of meeting time from Berlin, or
reduce it so that an additional hour(s) can be used for design team meetings
and other unofficial interaction?
Q3. Any other feedback? Other suggestions?
The IESG is looking to make a decision of the IETF-97 schedule by September 30,
so input before that is most useful,
Jari (for the IESG)