FOSDEM 2021 LLVM dev room?


LLVM has been present at FOSDEM with an LLVM dev room for many years.
The FOSDEM organizers have just put up a call for participation for the next FOSDEM, on February 6th and 7th 2021. It’s no surprise it’s an online-only event. See

In the past few years, I’ve been the main organizer of the LLVM dev room at FOSDEM, together with a number of other volunteers.
This year, it’s looking like I will not have the time available to be the main organizer on the day itself. Furthermore, with the conference moving to an online event, it looks like at least 10 volunteers are needed to be able to run the dev room, covering various roles as described below.

I’ve decided to not be the main organizer for the LLVM dev room at FOSDEM myself for the 2021 edition.
That being said, I am more than happy to support a new main organizer if someone steps up. If you’re interested, please let me know.
If you’re interested in any of the other volunteer support roles (described below), please let me know too so that we can have a guess of whether enough volunteers are available.
The deadline for proposing the dev room is really nearby: this coming Friday. So if you’re interested, please do reach out by Thursday at the latest.

Please note that it should be expected that the dev room will run roughly during day time Brussels time zone (CET).



PS. Please see the below notes from the FOSDEM organizers on the volunteer roles they expect are needed.

Our basic strategy is to replicate the various elements of the physical
event as closely as we can in an online environment. (Well, except
for the overcrowding!)

The main thing to appreciate is that the volunteer input required
to produce a professional online event will be significantly higher than
you may be used to. We expect a typical devroom to need at least 10
volunteers during the event so that people can take breaks.

Devroom managers will be responsible for finding and scheduling
volunteers to perform the various roles needed for an online conference.

Until the end of this year, the input required will be similar to normal
- selecting and scheduling talks. One thing to note here is that use of
our systems (e.g. pentabarf) will be compulsory and this includes calls
for papers. Systems will change and adapt and we need to be able to
contact all speakers directly whenever necessary.

The changes will be noticeable in January, when devroom managers will
need to assign specific volunteers to work 1:1 with individual speakers
to prepare recordings of their talks. All presentations will need to be
pre-recorded and put into our system at least a couple of weeks before
the event. We expect live presentations to be a rarity, but even if a
speaker intends to deliver their session live, there must be a recording
available to use as a fallback if something goes wrong and the live
presentation can’t be delivered for any reason.

As regards the features for devrooms, we are thinking along the
following lines:

- A main stream for each devroom.
Talks here will be pre-recorded, but questions will be taken live.

- A second stream for each devroom representing hallway discussions
that may follow each talk.

- A facility for people watching to submit questions.

- A facility for people watching to chat between themselves.

As such, we’re thinking of the following volunteer roles:

Prior to the event:
- A devroom manager and a deputy responsible for everything below.
This includes finding, assigning and scheduling all the volunteer
roles required.

- A programme committee to select and schedule the talks.
This is the same as a normal year.
Timeline: November/December

- Reviewers to help the speakers produce their pre-recorded content.
This is a new role.
Each speaker will need to be assigned to a reviewer who will be
responsible for ensuring the content is put into our system and ready
to be broadcast.
Timeline: January.

During the event itself:
- Stream controllers, one per stream, responsible for the content of
the outgoing stream at all times.
They will switch between inputs (live video rooms, recordings)
according to the schedule and intervene (sometimes on video) and make
decisions if there are problems.

- Chat moderators.
They will be responsible for monitoring the content of the chat and
dealing with any problems that arise (including banning people if
necessary). There might also be people assigned to answering
questions about the devroom content.

- Session moderators.
Every speaker will be assigned a moderator who will be responsible
for accompanying that speaker’s session, including all the preparation
and hallway chat afterwards, including checking the speaker’s video
connection is working, monitoring the chat and submitted questions
and asking them on video to the speaker.
The moderator will have the speaker’s private contact details and
make sure they have established contact with them about an hour
before their session.
A separate video room will be created for each session.
The moderator will meet the speaker (say) 15 minutes before the
pre-recorded talk is played, discuss the chat and selection of
questions privately with the speaker during the playback, perform the
live Q&A by asking the questions on video, then remain in the room for
the ‘hallway’ afterwards where they may, if they wish, invite other
people directly into the video chat for deeper discussions on the
hallway stream. Once the hallway broadcast ends, if participants want
to continue the conversation for longer, the video room will remain
available (but not streamed or recorded) and the session moderator
can close the room, or stay with it for longer, or hand over
moderation to the speaker or someone else they trust.

Ideally the Session moderator for any particular speaker would also be
the Reviewer in January so a relationship can be established.

Volunteers and speakers will also need to make time available prior to
the event to test that they can connect and use (with training) the
systems they will need according to their role(s).