Yep. We millenials spent decades talking about free and libre protocols (and software) and kids today love another walled garden against another one... good luck with that.

Inb4 "IRC sucks"... Jabber/XMPP exists since late 00's (at least ready enough compared to the first versions) and there are pretty fine clients for every OS.

Listen, I'm an old fart who may have been messing around on IRC when you were just a twinkle in your parents' eyes. IRC does suck along a lot of important metrics. The GPL open-source community-developed project I worked on for 19 years moved from IRC to Matrix several years ago, and the payoff in terms of engagement was obvious immediately.

I agree that walled gardens are a trap. But you're not going to convince people to move to free solutions without being able to recognize clearly why they walled gardens are so attractive in the first place.

I'm from 1987, are you sure? And I was talking about Jabber, not IRC.

> And I was talking about Jabber, not IRC.

Right, I misunderstood your last line. I initially took you to mean, "We've had IRC since forever and Jabber since the early 00's..." Reading it again, I now understand you to mean, "Before you say 'IRC sucks', which I agree with, better protocols like Jabber have been around since the early 00's."

No, I like IRC, but IM has different uses. IRC is for technical/non-private/random connections to roam around, like going a public place IRL. IM it's for personal stuff, as talking between relatives, close friends, word colleagues and so on.

IRC can be used for private communication, in fact it was the protocol for private communication used by for example warez scene groups years ago.

IRC, because of its simplicity, can be layered, so adding encryption is trivial if we want to hide from the server (FISH). By default IRC servers show hostname of the user, but that depends on the network; for example LinkNet hides it.

Right, well, the project needed something like IRC / Discord / Slack, where random people could just show up, ask questions, hang out, whatever. You know, engage with the community.

"Lots of clients" -- that's actually a problem. "Oh, you want to have a quick chat with experts in our community? First, here's a list of 50 IRC clients, half of which haven't been updated since 2003, all of which have different advantages and disadvantages. Go through and choose one and go through all the configuration. Oh, and you wanted to be able to read something someone wrote while your laptop lid was shut? Like, maybe you're not in the same timezone as many of the contributors? You're going to need to find an always-on server and set up this IRC bouncer. Also, because of spam, we only allow registered users to post, so you're going to need to type these runes in the command-line and make yourself a certificate."

Matrix certainly had its warts when we switched, but it was still an immediate quality-of-life improvement, particularly for newcomers.

I mean, HAM radio is basically a chat room with an entrance exam; there's certainly something nice about having communities full of people with that sort of filter up front. And, for a charity designed to teach kids to code, maybe helping them munge around with IRC clients is a bonus. But for a normal community, IRC isn't the right tool for the job.

> in terms of engagement

What's your definition of "engagement" here? Because it makes me think of social networking tactics to keep you ... well ... engaged ... the longest time possible.

I imagine that they mean engagement as in, "how many people in the company or group actually use the software on a regular basis".

Moved on to matrix? Many did... and they're all realizing matrix doesn't actually work long term. There's only the synapse server and there's literally no way to trim data from the db in synapse or everything breaks. That means the db just grows and grows until it's too expensive, or too slow (re:IO), to work. That's why the matrix.org homeserver has a 55TB db. That's why many long running IRC servers gave up on running matrix bridges because it simply became too computationally/resource costly to run the simple text based server even if they loved the features.

So unfortunately Matrix is a dead end. The matrix foundation gave up control 2 years ago. Matrix is now controlled by Element.io corporation and they only care about their government hosting contracts. It's really only viable if you have a significant constant money stream to pay for the ever increasing server resources like governments/corps.

IRC persists. It is the text chat layer of the internet which is the platform. Trying to build the entire internet into your text chat platform, and storing everything, is the kind of insanity only for-profit operations do... and eventually die from. Whereas IRC being a dumb pipe with lists of IPs associated with sockets will live forever. And cheaply.

> There's only the synapse server

https://matrix.org/ecosystem/servers/

> and there's literally no way to trim data from the db in synapse or everything breaks.

https://element-hq.github.io/synapse/latest/admin_api/purge_...

https://element-hq.github.io/synapse/latest/message_retentio...

https://github.com/matrix-org/rust-synapse-compress-state

Synapse server is the only server that does the things matrix says it can do. All those others don't. Notice only two of those say 'stable' and only one, synapse, implements the full set of features. Also can't switch between them.

And your links to synapse features ("Please note that, as this feature isn't part of the Matrix specification yet, the use of m.room.retention events for per-room retention policies is to be considered as experimental.") may describe certain synapse functionality but in practice is doesn't work and the db keeps growing as does IO load. The compression thing is an attempt at mitigation because the protocol just doesn't handle it.

is... was it Ellis island?

This. It is mind boggling to see an organization that teaches tech related stuff be so clueless about the dangers of proprietary software, cloud services and walled gardens.