Thanks for allowing me to introduce what’s left to do Jeff
That’s a very fair question, which requires a solid answer to be acceptable. The initial post of this thread is the first part of a series. I also blogged something about how to explain Matrix with e-mails, to make it easier to understand for people less familiar with the subject. In its conclusion I say
A post on how an organisation can keep its sovereignty in a decentralised and open network as Matrix will follow in the coming weeks.
That’s another very important aspect: Matrix is an open network, but it doesn’t mean we can do anything we want and expect it to work like a charm. The migration from Element’s to GNOME’s infrastructure is one of the steps of a broader plan. We need to think it through first, and to set-up a set of rules to enforce our strategy. My mind is pretty clear on that point, but rather than throwing ideas as I’m sipping my first coffee of the morning, I’d rather wait for a detailed blog post to explain it all in depth.
Then comes the last part, the one you mention: we have a Matrix instance in a given state, we will have a target, we need a transition plan to reach that target. By a happy coincidence, @averi is re-working our identities management system. Most of the steps on the Matrix side are clear to me, but instead of throwing ideas in a mess I would rather wait for our identities management system to be advanced enough to finish up and formalise the transition plan.
Matrix has pretty solid, battle-tested defence mechanisms now. The Matrix.org instance has more than 100k Monthly Active Users (MAU) and of course has triggered spammers’ interest. The most immediate defence mechanism is against individual troublemakers: a moderator in a room can kick/ban a user from a room, and redact their messages so they are removed from the timeline.
More sophisticated attacks can happen, for example when a spammer has automated the spamming in several rooms. That’s where the hammer of Matrix justice comes to play. To keep things short, it creates a room in which a moderation team can send commands to a bot to ask it to kick/ban a user and redact their messages for all the rooms under its protection. Just a note: it doesn’t make sense in the Matrix universe to talk about “rooms belonging to a server”. What makes more sense is to talk about “rooms we control”: that’s why the moderation by mjolnir is opt-in. When people create a room and want to be protected by mjolnir, they need to grant it the administrator privileges to be placed under its protection.
Finally, for the more extreme cases there are two last resort mechanisms: putting an obviously maliciously crafted instance on a denylist, and for the most extreme cases shadow banning them.
That’s where things get really good: with the proper authorizations on the IRC side it is possible to kick and ban users from the Matrix side! For more sophisticated defence such as changing the modes of a channel, manual work has to be done on IRC. We’re still working with GIMPnet administrators to upgrade the network, and we’ll be able to retrieve proper rights to do it.