[Galene] About subgroups

[Galene] About subgroups

[Juliusz Chroboczek]

Hi,

I've recently had a telephone conversation with a user who told me that
none of his colleagues understand how to use the auto-subgroups
functionality in Galene.  Since I'll be writing it down for them, I might
as well share with the list.

The auto-subgroups functionality was designed for CS practicals, where 12
to 30 students are programming in a relatively autonomous manner, with
occasional help from a lecturer.  The practicals are usually not graded,
collaboration between students is encouraged, and the students are
provided with a detailed list of steps.  Here's an example:

  https://www.irif.fr/~jch/enseignement/sieci/lab4.pdf

Before the beginning of the semester, I create a Galene group called
something like

  https://galene.org:8443/groups/advanced-network-programming/

with "auto-subgrops" set to true.  For the lecture, I arrive on time,
about half of the expected students are already logged in.  I ask the
students who have a cat to switch their camera on and show us their cat;
this creates a relaxed atmosphere, but also makes students less
self-conscious about participating (nobody wants to be the first to switch
their camera on).  When the students start joking among themselves in the
public chat, it is time to start.

I ask the students to switch their cameras off, and to only switch them on
when they have a question to ask.  During the lecture, I keep my camera
on, I share my slides, a drawing appliation for impromptu diagrams, and
sometimes a terminal to make a demo.

When it's time to do a practical, I tell the students to organise
themselves into teams of 1 to 3 people, and to pick a nickname for the
team.  I then ask each student to remain logged into the main group, and
to also join a group called

  https://galene.org:8443/groups/advanced-network-programming/team-name/

In that subgroup, they are required to share their editor window, and
I tell them it's okay to use it for chatting with their team members.
Most of them switch their microphone on, some of them also turn the camera
on (usually the ones with cats).  Some groups want more privacy, and they
chat on a channel I don't control (usualy Discord, it gars me greet).

I then switch my camera off, I make myself a coffee (instant, no time for
proper brewing) while the students get organised, and come back to the
main group.  I type

    /subgroups

which gives me the list of teams.  I then join all of the subgroups, and
mute all the browser tabs except the one with the main group.

I mostly look at the main group, where students ask questions in the chat.
If a question is easy, I simply answer in writing in the main group; if
a question is more involved, I ask the student which subgroup they're in,
unmute the subgroup, and discuss the issue together with the team.  If
multiple students have asked the same question, I switch my camera on in
the main group, and explain the issue to everyone.

If the main group is idle, then either the students are working (the
practical was written well, they don't need my help), or everyone is
sleeping.  When that happens, I roam around the subgroups to see what
they're doing (that's why sharing the editor window is compulsory), which
gives me an idea whether I should leave them in peace or intervene to wake
them up.

This procedure requires juggling with anywhere between 5 and 20 groups,
which is impossible with any software other than Galene.  With a little
practice, it works fairly well: I've managed to lead successful practicals
online during Covid and, more recently, when I couldn't do my practicals
in person, which, false modesty put aside, I'm rather proud of.

As with all things Galene, the functionality is there, but the UI needs
improving.  I need to make the output of the "/subgroups" command
clickable (an easy fix, just not something I've gotten around to yet).
There should be a way to find out which subgroup a student is in without
asking them, but that's made somewhat difficult by the lack of user
identity in Galene.  (Galene handles authentication, not user identity --
there's no way to determine if two users logged with the same username are
the same person.  That's a feature, by the way, not a bug.)

Your usages are likely to differ, please feel free to suggest improvements,
as long as they don't harm my use case.

-- Juliusz