IRC is the only viable chat protocol (2022)

IRC has some advantages. You can use with without specialized software (and I have done many times when not having a IRC client), or you can write your own IRC client or use one of many. There may also be web interface if you want them.

Note that you can make IRC server which makes logs of channels and could provide a link to access the file in the MOTD or HELP file, if desired. It can also be done by clients and link in the TOPIC message, which is more common.

IRC is good for fast communication with short messages, but for slow communication with long form messages is also sometimes more useful and for that, NNTP would be better, I think.

That is to say, Discord's search actually works. It may not work great, but it will find all posts containing the word "dog" when you search for it. The search in windows explorer, on reddit and on most forums will often fail to find an item even if you search for its exact name. Last I tried it, windows explorer couldn't find a file in the folder I was already viewing. It's impressive.

fwiw, i would be first to agree (as project lead for Matrix) that we (I) have fucked up the communication about libera irc bridging problems, and I am sorry for it.

the root cause is lack of bandwidth: the bridge is both written and run by one guy, alongside a tonne of other stuff which often takes priority given it revolves around trying to generate $ to fund matrix dev.

Around the beginning of the year, we had some nasty problems on the bridge (from memory, some delayed traffic, and a security bug). Libera started threatening to deportal us, so we rushed out a major feature (persistent IRC connections, so Matrix users don’t reconnect when the bridge restarts) to try to address the longest running series of issues and avoid Libera killing portals (ie stopping users bouncing via Matrix). Unfortunately, while this seemingly worked, it introduced the nightmare subtle traffic loss bug which the parent is complaining about, which we haven’t been able to repro outside of production. Meanwhile, rather than averting deportalling, the new problem reinforced the case, combined with false concerns about archive.matrix.org (https://matrix.org/blog/2023/07/what-happened-with-the-archi...).

This has caused a particularly unpleasant vicious negative feedback loop where we’ve had to split time fixing the reliability bug with: fixing plumbing, preparing deportalling, handling tonnes of community reqs about deportaling and the awkward UX of plumbing, dealing with the libera team, and dealing with a new brace of security issues which showed up thanks to newfound attention on the bridge.

Meanwhile it’s fair to say that we have badly failed to communicate about this with the wider IRC community, focusing on trying to fix bugs while talking to the libera team instead. This is clearly a huge screwup, and, again, I apologise for it (and for the bad job I’ve done on resourcing keeping the bridge running).

At this point, the our best hope is that we’ve fixed the root message loss bug, and that the bridge will go back to being lower maintenance again, and plumbing will prove usable in practice. However, we’re also in a catch 22 when the main use of the bridge (portalling, ie acting as a bouncer) is now gone, meaning fewer people will be able to use it and it’ll be even harder to justify spend time on it going forwards. If libera hoped that by turning off portals they would somehow make it easier for us to improve the bridge, they were wrong.

Our fallback position, if plumbing doesn’t work out, is to try to find a set of people who libera trust to run/maintain the bridge and hand it over to them - or alternatively give up and encourage the community to run their own personal bridges via matrix-appservice-irc or heisenbridge or even matterbridge, if empirically running a huge bridge to libera’s satisfaction is impossible.

If anyone has any other bright ideas, please let me know.

Finally: the reason we have run the bridge for the last 8 years is, perversely, to try to support IRC and let xkcd.com/1810 style users still participate in communities which have otherwise migrated to discord/slack/matrix and avoid community fragmentation. Right now the Libera Matrix bridge has around 10K IRC connections - around 25% of Libera’s total users. By knifing portaling, I’d guess this will drop 10x. So the whole thing ends up being an utterly depressing spiral of fail.

You don't have to take responsibility for fixing something just because you criticized it.

Fixating on the term used is a tad disingenuous. I think most people reading hacker news understands that the IRC ecosystem is being referenced here - not the protocol itself. IRC has been around 30+ years and it still doesn't have any great clients.

If you wanted to build a featureful chat client IRC wouldn't even be a good choice because there is so many things missing from the spec.

discord makes hosting a real-time online place for most communities of all sizes very easy. it also doubles as a free cross-platform messaging app with a massive userbase. as the top-level GP posted, it mostly has a superset of the features of other community/messaging platforms (at the cost of searchability and data ownership/privacy).

as a student, I'm in discord servers for friend groups, school clubs, unofficial classroom servers, game communities, fanbases, etc - anything that can benefit from conversations organized into different channels is usually turned into a discord server.

when I think "how can I take this group of people interested in the same thing to gather together online?", the alternatives that come up are:

- facebook (too linked to real life identity, also not really great for separating conversations into specific channels to my understanding) - reddit/forums (which can compliment a discord server in being more searchable/async, but in no way rival Discord's real time text/voice chat) - mumble/teamspeak/skype/matrix (largely just has a subset of discord features) - slack/microsoft teams (pretty much solely designated for work and not at all marketed as a general purpose communication solution)