Ergo Chat – A modern IRC server written in Go
github.com
github.com
The v3 chathistory support and the always-on[1] multi-client[2] features paired with modern clients (like Goguma) go a long way at providing a modern chat environment. Most others on the server don't even know that they are using IRC.
The built-in websocket support is another key feature for me since it lets me to provide a web client just by serving some static files (I use Gamja for this).
[1] always-on: https://github.com/ergochat/ergo/blob/master/docs/USERGUIDE....
[2] multiclient: https://github.com/ergochat/ergo/blob/master/docs/USERGUIDE....
Pre-LLM-GenAI gathering information from huge chat logs was quite limited.
But different extraction mechanisms always existed. Some people kept logs, some servers/channels had web-archives.
Discord tries to keep it exclusive to them.
(Focussing on technical side here, whether it's socially wanted is a different big question, which ends with a "it depends")
grep/recoll. Fast even on ancient machines.With discord the message might get pinned if a mod thinks it was useful and remains undiscoverable for the majority of people
IME it doesn't work very well. If you hang around a discord server long enough you see the same conversations happening over and over and over. And the app is not good at helping you finding old conversations (which is why they are repeated endlessly).
It's also very bad at resuming where you left off, so it's incredibly difficult to follow an active thread without missing anything:
https://www.reddit.com/r/discordapp/comments/18d4tiv/inexpli...
You can also log from the server; using a multitude of modules; https://docs.inspircd.org/4/modules/log_syslog/ or https://docs.inspircd.org/4/modules/log_json/ if you're using inspircd.
Turns out what matters isn’t how easy it is to access logs but whether anyone cares to do it.
Heres an archive searcher for #bitfighter on freenode: https://bitfighter.org/irclogs/search.php
A channel with one day logged in 2024, then nothing for two years, then nine days logged in 2022, five days logged in 2021, etc?
What about big tech channels like, say, nodejs and javascript?
The point that I am making is that the capability and software exists, even for others to view, if you want. I mean: did you need an account or specialised client for these logs?
Thats the point.
Be the change you want to see if you want archives uploaded. Or, do what others do: grep your local files or znc logs.
And for IRC, I think this is a good compromise between "everything is public at all times" and "everything is walled off and private at all times".
If you want to have a log of your use, it's possible. If the org running the server wants a log, they can do that too. Is that possible with something like Discord?
So…
IRC is an open, community protocol, and as such not a walled garden. Even if I'm not involved, which I could be, I trust the composition of people there much more than any single commercial actor. The power dynamics are fundamentally different.
A single instance (an IRC network) may be a "walled garden", in control of the group that runs it. The incentives are different. Also because people can simply migrate to another network given the open protocol (and different, third party clients; the API cannot be shut down like with Twitter/X). Historical example e.g. Freenode ownership change.
IRC is not a walled garden. You will have a better time registering with Nickserv, but you can pull content and logs channels. So, more like a fenced meadow.
Central Login Server hosted by Foundation of Chat Program and servers authenticate via OAuth2.
Central Login Server keeps track of which Servers identity is in and their endpoints so people don't lose track of servers.
Central Mobile Notification Service because mobile app notification is so important AND SO PAINFUL TO RUN.
Federation NOT required.
Even getting people to use a different messaging app (such as Facebook Messenger) already installed on their phones is difficult.
I have whatsapp and conversations and that's it. I only use teams on my professional computer, you will never see me install "work" on my personal smartphone and nobody needs a linkedin app, the website is enough.
I agree about LinkedIn and will probably get rid of it; I mostly installed it as I was going to a conference and wanted the easy option for swapping contact details using camera codes.
Probably that they don't bother to have another chat app just for a few people.
A older family member has WhatsUp and Signal and gets sometimes confused. I have WhatsUp, signal, telegram, discord, meta messenger and would be .. hesitant to install an IRC client for like 3 people.
And lose because you can't give it to a kid that doesn't have a mobile phone number.
I have shared custody of my daughter and we communicate via xmpp on a tablet they carry over there when they spend a week at their mother's.
Want to find that recipe that you remember that some person shared with you a while back? Now you have to look through four different apps.
And the overhead grows quite fast as you add apps.
If self-hosting in general wasn't such a PITA, I'd probably research the options and set something up. But honestly, I'm burnt out with trying to maintain even the most basic setups. I have a Raspberry Pi with NixOS under my desk that hosts Miniflux over Tailscale, and I can forget it exists 99.7% of the time - until I accidentally unplug it, and 6 hours later, wonder wtf happened again.
Now multiply the problem by the average funny cat video size and crappiness of my residential uplink. Won't happen.
This is a great feature for privacy though
This “limitation” is the ultimate advantage from the perspective of Signal’s core competency.
Then what if a new person joined the group convo? Do they have a right to see everything that everyone has said in that group convo, right back to the beginning? What if someone objects to sharing past conversations? What if sharing past conversations is a legitimate security concern to the group?
How does one filter out those past statements by someone who doesn’t want to share past statements? With security-state and foreign-state moles being a rather big issue in some groups, this is a legitimate rabbit’s hole that needs addressing. Some companies that standardize on Signal may not want any prior convos to become available to new entrants.
Personally, I see the lack of history to be a very real competitive advantage, and not any sort of a nerf.
Sure, your own history can be shared between installs on devices that you yourself own. But that is your chat history, meant for only you.
When you add a new computer, Signal refuses to sync the history.
Where I live (Canada) only a single person I know uses Signal. Everyone else is on Whatsapp and/or iMessage. As far as I am concerned, Signal is a wasteland.
I have received 400% more spam on Signal than I have received real messages.
Is everything related to potatoes lol?
I didn't bother setting up proper push notification so we use the "Servers supporting chathistory" mode. This means that when the app is not in the foreground a workmanager task polls every few minutes. So in this mode notifications can be delayed by a few minutes which is fine for our use case.
You can run something like thelounge and have that on any server.
We tried to offer a bouncer instance for users and even that had a barrier to entry because it requires creating 2 accounts: one for libera and one for pico.
I think about us switching to ergo every few months because I think the onboarding experience will be much nicer.
Logging into a channel for the first time and see the chat log will make people a lot more motivated to stay.
Once a server starts hosting for people there are a huge set of problems and pressures. IRC is the text chat layer of the internet. It's not by itself, it's part of a larger network of open protocols. If you avail yourself to these it far exceeds the capabilities of even the most featureful walled corporate garden.
My little IRC community has been going on for what must be more than 15 years now. There has been a noticeable slowdown this year. Maybe it will be the end of the line in a few years.
Nice to hear that at least there are others out there. I have vivid memories of the first time I connected my Nokia to a laptop and used mobile dialup-internet from the passenger seat of a car to connect to the IRC server while going down the Autobahn. Mobile internet was unheard of at the time, so it was totally exciting!
:User1 PRIVMSG User2 :Hello, are you receiving this message?
My memories are from a long time ago so I may be overreacting... perhaps the Ergo authors can comment on their experiences if they are around here! I heard about IRCv3 but I doubt that effort solved most of my main gripes with the protocol.
If I were to work on a messaging app today, I'd look elsewhere for inspiration. From a quick search, it seems there's room for a modern and _simple_ protocol for chat, simpler than XMPP or Matrix. Essentially, we need a protocol that is for messaging what Gemini is for HTTP. Stretch: squinting a bit, the NATS client protocol looks close to a starting point for something like that [1].
--
1: https://docs.nats.io/reference/reference-protocols/nats-prot...
Are we once again at a point where it's time to start over?
A federated asynchronous group chat protocol with modern e2e encryption for desktop and mobile use (thus, not always online) is impossible to build without a faire share of complicated corner cases.
Ask Sisyphus.
In the meantime, while everything else keeps going through newer iterations of XKCD 927 (https://xkcd.com/927/), IRC keeps chugging along, and gaining new features and functionality as it enters its fourth decade despite the previous commenter's gripes.
that doesn't make any sense.
A bunch of people here talk about setting up a small server for family and friends, so I think a smaller protocol, akin to Gemini in size, could be a lot of fun to work with for small scale deployments.
There may be some scalability differences between different protocols/implementations for the admin of the service, but Snikket fits comfortably on even low-end Raspberry Pi devices, and literally over half of the typical resource usage is by the web dashboard (yay Python).
So what difference does the protocol make? It can make a difference to the developer experience. If all you want to do is exchange text messages, then yeah, XMPP and Matrix are absolutely overkill. But - especially for a family-and-friends use case - people also want file sharing, audio/video calls, and all that stuff. It very quickly gets quite complex to support all this stuff, especially in a way that allows you to evolve the protocol over time (trust me, what you think of as core messaging features today, were not a thing 10+ years ago, and messaging in 10+ years will also involve a new set of features).
There will always be a set of users for whom plain text messaging is enough (90% of my own daily communication is via messaging in a terminal app). However that set does not intersect significantly with the general population, and practically none of my family members would accept such a solution as a replacement for WhatsApp.
One question: are you aware of Jami[1], f.k.a. Ring? If so, how does it compare to Snikket?
I see that Snikket requires a server, whereas Jami is P2P. The benefit of a server is probably that messages can be stored centrally and not on each device. But I can see pros and cons of either approach.
[1]: https://jami.net/
You're right that there are pros and cons. Obviously, not having to run a server is a big pro for many. However, the first thing to remember when researching messaging solutions - no matter what anyone tells you - there are always servers! What differs between projects/platforms is what the servers do, and who runs them.
Jami uses a network of public servers that form a distributed hash table (see https://github.com/savoirfairelinux/opendht ). It's a neat design, and they have done a good job tackling the challenges of P2P messaging. Last time I looked in, it still required both users to be connected at the same time for message delivery/sync to work (the devices use the DHT to discover each other, and then exchange messages). This is a fairly common issue for P2P systems, and can be frustrating in a mobile-dominated world. Their DHT software does support push notifications, which helps with that though.
Another project in this category is Briar, which uses the existing network of Tor servers - and therefore adds IP address masking and a layer of metadata protection (as always, there are limitations, e.g. https://code.briarproject.org/briar/briar/-/wikis/FAQ#does-b... ). Briar built the concept of "mailbox" nodes you can run ( https://briarproject.org/download-briar-mailbox/ ) to overcome some of the problems with P2P messaging.
With Snikket, instead of using existing publicly shared infrastructure, you just run your own server (e.g. VPS or Raspberry Pi or whatever) which is responsible just for your users, and your users connect directly to it, improving (meta)data locality. This makes the design very simple, reliable and efficient (e.g. with battery/bandwidth). It also enables some important (for our use case) convenience/UX features, such as the ability to add restrictions on certain accounts (e.g. for children), and server-managed contact lists so all your family members don't have to manually add each other as contacts one-by-one. Things like that.
No approach is universally better than every other, but I much prefer the Snikket model for the family-and-friends use case. Not that we don't have our own challenges. Our iOS app is probably the weakest part right now (in terms of UX and general polish). Something I'm working hard to get fixed in 2025.
Yeah, there are definite challenges of the P2P architecture. But like you say, Jami seems to have done a good job addressing them.
I looked at Briar, but it has a different focus and is more limited in functionality and less polished than Jami. My use case is text messaging and audio/video calls with a close group of contacts, so Jami and your project look like a better fit. I also considered Matrix/Element/FluffyChat, but the Matrix architecture is confusing, and the clients are underwhelming.
Anyway, good luck with Snikket! If Jami doesn't work out for me, I'll definitely give it a try.
Matrix has been all about superlative marketing and over-promises with little substance and theoretical underpinnings to back it up.
The fact that they were spreading FUD and disinformation against XMPP and the alternatives in their early days should have been a red herring. The fact that they haven't stabilized their protocol and made it scalable to mid-level deployments after more than a decade should be a reason for pause. The fact that after so long, due to the sheer complexity of it, only one entity has emerged as willing to sink the costs of Matrix should be a reason for worries.
I think that we are due for an XMPP resurgence. Not that it is perfect by any stretch of mind, but it is no-BS, mature, lightweight on resources and bandwidth and verifiably fit for purpose, whether you want to host small-scale for your family & friends, or for the whole world with the ambition to be the next GTalk, facebook messenger, or whatsapp (all of them XMPP services at one point or another of their histories)
I think you meant to say red flag, not red herring: https://en.wikipedia.org/wiki/Red_herring
I don't think so. The underpinnings were there, may it be the concepts for E2EE, P2P, Verification, Bridging, decentralized rooms and conferencing, synchronized history, etc. Not everything is production ready let alone finished with perfect UX, but it doesn't lack theoretical underpinnings.
> The fact that they were spreading FUD and disinformation against XMPP
In the past, I have experienced it more often vice versa than this way.
> The fact that they haven't stabilized their protocol and made it scalable to mid-level deployments after more than a decade should be a reason for pause.
There are already pretty large deployments (1+ million users), so I think it's scalable to mid-level deployments.
> only one entity has emerged as willing to sink the costs of Matrix should be a reason for worries.
idk what you mean
> all of them XMPP services at one point or another of their histories
histories is an important keyword
Let me cut it short: either you think that there's nothing to be said about the current state resolution machinery, and that's an admission that Matrix will never reach its stated goal to become a mainstream decentralized network (because federation at scale just doesn't work with the current status-quo), or you agree with me that something needs to be done (but, like me, could turn impatient that nothing materializes after a decade).
>> The fact that they were spreading FUD and disinformation against XMPP
> In the past, I have experienced it more often vice versa than this way.
I am not challenging your experience, I am talking about what was effectively written about XMPP on the Matrix website. If this was reciprocal as you say, I'd like to be shown where and in which terms XMPP has engaged in disinformation campaigns against Matrix. At least, with the xmpp.org website's history being on github, it should be very easy: https://github.com/search?q=repo%3Axsf%2Fxmpp.org%20matrix&t...
>> all of them XMPP services at one point or another of their histories
> histories is an important keyword
If by that you mean that WhatsApp, the single largest chat network to have ever existed, is history, I've got news for you. While we may regret that federation never was a thing with this network, talking about underpinnings, they are still very much versed in XMPP. And that's not an exception. All Android devices depend on XMPP for push notifications, so do every Nintendo switch, and millions of access points, networking and IoT devices. XMPP is all but dead.
If folks are interested seeing the companies currently sinking $ into Matrix by routing $ to the Matrix Foundation, you can see them at https://matrix.org/support. Thankfully it’s not just Element these days, but companies like Thales and Huawei too. That said, there’s still a $ gap in funding.
Meanwhile Matrix 2.0 is a genuine pleasure to use, imo. You can see the substance and explore the theoretical underpinnings at https://matrix.org/blog/2023/09/matrix-2-0/ and https://element.io/blog/experimenting-with-matrix-2-0-using-...
I do regret saying that XMPP had fallen behind, 10 years ago, when we launched Matrix, though - not least because 10 years later folks are still bringing it up ;)
"trust us it's good this time" → (still the very same at its core) → "you are right, we are sorry, but bear with us, <buzzword> is just about to be ready!" → (buzzword is abandoned/releases but doesn't improve things in a meaningful way) → "trust is it's good this time" → …
Matrix 2.0 on the server is all about "sliding sync", which, once you read through the buzzword, is about running the client on the server so it sucks less at fetching history. Even that works super randomly, if it works at all, even on EMS-hosted instances. The rest is the same. On the client-side, Element X is not daily-drivable because of the bugs and missing essential features. Nothing has improved in a year on that front, which is reminiscent of dendrite: dropping it for the sake of salvaging the original hack would surprise nobody at this point.
> I do regret saying that XMPP had fallen behind, 10 years ago
By no means am I saying that XMPP was great 10 years ago (what was back then, really?), but that's a very poor rationale for torpedoing another project by spreading lies about it on your FAQ, especially when that project was working on the problems and eventually came to address them, across a whole ecosystem of clients and server, while Matrix to this day is still poking at the beast wondering how to tame its single-vendor/single-client/single-server complexity.
It might be worth considering that whining about Matrix is misdirected anger - it may be better off pointed at the proprietary centralised surveillance capitalism alternatives rather than us…
Outright denying people's bad experience with Matrix has also been a sad part in Matrix marketing for so long. Unfortunately, I don't have experience with Element X, since my provider doesn't support it yet, so I can't confirm or deny that testimonial, but my exceptations are not high.
> It might be worth considering that whining about Matrix is misdirected anger - it may be better off pointed at the proprietary centralised surveillance capitalism alternatives rather than us…
It's not either or. People are frustrated on the other hand because of the annoying proprietary platforms, but also because there are currently no alternatives to them that one could whole-heartedly recommend as a replacement.
I think if you look at my comments on HN you'll see me spending most of my time trying to explain or apologise for people's bad experience, but ymmv.
> Unfortunately, I don't have experience with Element X, since my provider doesn't support it yet, so I can't confirm or deny that testimonial, but my exceptations are not high.
Shame, you're missing out.
NATS as an IRC replacement is using the wrong tool for the wrong job.
I've long been thinking about how I'd replace IRC and I keep coming back to another IRC-era protocol that wasn't as popular but was definitely ahead of its time by some amount: Hotline. It wasn't Open like IRC and was limited to Macintosh users for a long time, but it has a lot of what most people are looking for in a "chat system" today. One thing, however, that I think would benefit the protocol designers of the future: Learn from MSN.
MSN separated their Authentication and Chat systems. You ask for a token that identifies you to the chat system from the authentication system. These were two separated protocols that did one thing specifically for each other. I honestly think that separating chat-locus and identity from one another is important.
But from my recollection of writing bad IRC client in the 90s (including in mIRC script after they added support for sockets in the scripting language in mIRC v5.3 released December 1997), the protocol isn't that hard to work with. Once you figured out you needed to send a : in front of your message text, I don't remember any other big gotchas.
It's not XML, and I don't remember any quoting issues, but maybe there's some I forgot. CTCP maybe isn't ideal either. DCC probably doesn't work anymore, but it was a different time when being on the internet meant an expectation that you were a peer and could accept inbound connections.
You could write your own servers and services too :). My mIRC scripts all had s:lines on my network, all running in my mIRC (NoNameScript) environment, which I'm sure were on Hawkee somewhere.
Complaining about IRC in 2024 is bleh, by the way. The protocol is what it is, you've had a long time to get over it and if you were late to the party you need to realize that yes, there are tons of old protocols that are pretty badly designed, welcome to the world. IRC as a way of public chatting (and maybe even more so running a closed chat) is still somehow among the best we have and personally I'd prefer it if Slack, Microsoft Teams and Discord died completely for "Hey, we've set up a place to talk about this work/event/hobby thing!".
I recently went through the hassle of deciding on something small for my family + company circle. Mainly considered XMPP and Matrix, and went with Matrix. Didn't know there was such a thing as "always-on" on IRC tho.
1. The only stable and maintained implementation is "matrix-synapse" and it is written in Python.
2. The most commonly used client is "element", and it is governed by the same people. So it feels we are the mercy of a single company.
I wanted hard to go with a more established protocol like XMPP but failed to get a server running properly :)
That IMAP auth trick is really awesome thinking BTW, kudos!
Dovecot is really great, and a ton of stuff supports using it as a sasl auth backend (postfix being an important one). I made a simple facade service that feeds it and postfix from couchdb via its dict backend[0] and postfix's tcp_tables[1], then point everything at dovecot for auth. Couch document IDs map really well to email/user, domain, and sieve script lookups; helluva lot simpler than setting up and managing LDAP.
If you want to selfhost just make it easy and get Conduit. Its single binary and uses embedded db (rocksdb or sqlite). I cant say about federation but for private chat server this has been solid for me for years. I still run it with sqlite (worse than rocksdb) and with 30 very active people its more responsive than Synapse ever was.
curl example.com/.well-known/matrix/client
responded with an incorrect base_url.
It's a server feature and might be unique to Ergo. It's a per-user setting (with a global default) and when it's on the user always appears to be online so you can type at them any time like you would with other DM systems. The v3 chathistory support ensures that they don't miss messages. For clients that don't support chathistory the server can replay any unseen messages.
It's a lot like what bouncers provide but integrated and very easy to enable for everybody so no extra steps required for my users.
Document your shit in proper places.
I also find it annoying that the creators of slack pretend they invented something original.
Without a server-side backlog, the chat is fleeting, and everyone knows that, so to preserve important content, people know to save it somewhere else. This keeps the chat as it was meant to; a live chat, mimicking that of a human conservation, where nothing is recorded until someone makes the conscientious effort to do so.
Every single community has become a silo with their own memes channel. It's like they emulate modern social media websites where they try to keep you engaged and in one spot forever.
I find it dystopian and power-tripping personalities trying to invent rules upon rules on their little kingdom is really not conducive to spontaneous socialisation.
Now, sorry, you cannot contribute to this conversation because you haven't fully read the rules on #welcome, didn't complete the captcha from our bot and, worse of all, you have not chosen a role for yourself.
Software shapes behaviour. If you design software for social interaction you are designing mass behaviours.
These forum channels don't automatically appear in the sidebar and you get no indication of anything new added unless you specifically follow the thread.
Moving into forum threads is a great way to kill a conversation, perfect even, honestly I wouldn't be surprised if the feature was added just so Discord could appear in "forum" searches. There's no way the feature was planned when it's so bad for conversations.
In any case the discord mod quirk of stomping on discussion inertia plus this horrible new forum system guarantees a conversation fizzles out immediately.
Thinking about it that use case makes far more sense than what the communities I'm in are using it for (basically just dynamic #channels that anyone can create) - especially with how nobody is a part of the discussion by default. That does make way more sense for a helpdesk kinda setup rather than misc discussion threads.
Seems like a neglected area of research in UX/UI - perhaps an opportunity for a good PhD or Masters: To what extent do software paradigms enable/inhibit pathological/virtuous personality traits?
And no - not all software is a "neutral" blank canvas. Behaviours cluster around latent or implied structures. Designers imprint their own values on code, perhaps even unconsciously.
Discord servers can have forums like channels. https://support.discord.com/hc/en-us/articles/6208479917079-...
The irony of the persistent backlogs for Discord means that a lot of this information is practically lost. Archive.org cannot preserve copies, so if the Discord server is closed or whenever Discord itself decides to call it quits, it's all gone.
The AI could maintain a Wiki with relevant stuff gathered from the discussions.
Not sure if it's good enough today, but it does seem like an interesting prospect. Not many people like filling Wikis.