One remark: I suggest to replace port 5223 with 443. It improves reachability in badly configured public wifis, such as some British trains.
this post was submitted on 06 Oct 2025
41 points (100.0% liked)
XMPP
485 readers
13 users here now
XMPP (aka Jabber) is the community-owned standard for real-time federated messaging.
JoinJabber.org admin support chat
founded 2 years ago
MODERATORS
That would be wonderful, indeed, but it is work the client developers need to do and it might take some years to get there.
A server admin probably wants the highest possible reachability with existing clients *now*.
I.e. either ALPN (which is easy with nginx https://wiki.xmpp.org/web/Tech/_pages/XEP-0368#nginx) or having a spare IP for xmpps.
I run OpenFire (https://www.igniterealtime.org/projects/openfire/) as my server. Upgrades are easy, and I like the 2000-era interface as it doesn't take up much room, but if I was starting from scratch again, maybe I would try a different server, as the way it deals with Lets Encrypt certs is a bit annoying when it's time to renew - this could be user error, but it's not obvious.
ejabberd having it's built-in "certbot" equivalent sounds great!
I use Monocles (Android) and Dino (Linux) clients, and it's pretty stable. Monocles also allows some server commands from the client, which is pretty cool. Sorting out TURN was a faff, but it doesn't help that my server is behind 1:1 NAT I guess :/
Still, now that is sorted, it's great :) I just need to find out a way of enticing my friends to use this instead of WhatsApp etc
Glad to see I'm not the only one saying that.
Just another data point: I am a millenial and none of my non-tech friends care about reactions; my parents and other relatives of their generation have no idea about how to use them, even in WhatsApp (which in my country has a near-monopoly on instant messaging). The only category of people I know IRL who care about the "outdated look" of XMPP clients is made of my "I think I know a lot about computers because I'm a gamer and I use SuperOptimisator 3000 to get 500 FPS in my games" friends. 🤷
Your experience is different, fine. Matrix has other strenghts, fine. But I don't understand the need for matrix users to come shit on every remotely popular XMPP online post everywhere, with what's basically FUD. In another comment, you said there is neither "emoji" or "reaction" support on iOS. WTF. While reactions are not (yet) implemented in Siskin or Monal, there is absolutely nothing preventing you from sending and receiving emojis in those iOS clients. Some people enjoy DeltaChat. I don't see them coming and shitting on every popular XMPP or Matrix post. Why is this like this with Matrix?
FWIW, I would be happy if everyone moved to Matrix instead of the privacy nightmare of walled gardens that are the de facto norm. I've started hosting my XMPP server on the cheapest VPS I could find about 10 years ago, and I have been happy with it. My parents use it. I am probably not going to retire that VPS anytime soon.
I don't know why you keep harping on that point. Yes the XMPP iOS client situation isn't ideal, but on all other platforms XMPP clients are significantly better than Matrix clients and outside of the US and Japan, hardly anyone uses iOS. Besides the fact that Apple is extremely hostile to open-source developers.
At this point I almost think not supporting crappy Apple products is a feature 🤷
hardly anyone uses iOS.
I keep harping on this point because you are doing wishful thinking. The numbers for iOS usage in Western Europe are around 32%. That's far from "hardly anyone". In the UK is over 50%, in Australia it's over 60%.
If you care about keeping in contact with friends and family, you'll be bound to find someone on iOS. What are you going to do, tell them to get a new phone just so they can participate in the family group chat?
At this point I almost think not supporting crappy Apple products is a feature.
Yeah, those grapes really are sour...
Mostly company phones in Europe and I suspect elsewhere too. Private use is very rare, and mostly old hand down phones from someone that got one from a company. And besides... nice cherry picking of numbers with "western" Europe only...
And no, I would tell them to install Monal. It works fine for normal messaging. But I never had the problem, because I know almost no one that uses iOS privately 🤷
Mostly company phones in Europe
Does it matter? It's not like people everyone carries two phones around.
and I suspect elsewhere too.
Wishful thinking, again.
nice cherry picking of numbers with “western” Europe only…
The point is that there is a non-negligible number of people using iOS everywhere. In Brazil, it's 17%, In China, it's ~20%. In Russia, 30%. Saying that "no one uses it outside of the US and Japan" is absurd.
I would tell them to install Monal. It works fine for normal messaging.
If you "know almost no one that uses iOS privately", you are probably surrounded by people who think that "emoji support" and "reactions" are not important. Try stepping out of this bubble and see how far your can go with Monal. People expect a lot more than "normal messaging".
And no, I'm not saying this out of hate or because I prefer Matrix. I wish XMPP could be a viable alternative, just like I wish less people were locked on iOS. But no amount of wishing will change the reality. If XMPP enthusiasts really want to make it succeed then unfortunately they need to be ready for an uphill battle and be willing to spend resources to get a compelling alternative working on a hostile platform.
Most companies prohibit the installation of non-company approved apps, so yes you need to carry two phones.
I am not going to lose sleep over XMPP having less than ideal support on a platform that is hardly used by the realistic target group of XMPP. And reactions and emojis work fine on Android and other platforms 🤷 Afaik emoji reactions are also going to be supported by Monal soon, but then there will be for sure some other thing you come up with why XMPP can't ever work 🤦
Plus, I think you have a totally false problem analysis if you think the less than ideal iOS support it the thing why XMPP isn't as popular as Signal for example.
I think you have a totally false problem analysis if you think the less than ideal iOS support it the thing why XMPP isn’t as popular as Signal for example.
I never said it was the thing, and you are creating a strawman by implying that I'm expecting it be "as popular as Signal".
What I am saying, plain and simple, is "XMPP might be technically superior in every possible aspect, but this means absolutely jack shit if most if not all of my circles have at 10/20/30% people who will not use it because the client is so feature poor."
The numbers for iOS usage in Western Europe are around 32%. That's far from "hardly anyone". In the UK is over 50%, in Australia it's over 60%.
When you move the goalposts, please fill in the old holes. Thanks.
What goalposts are being moved?
I think something to keep in mind with that issue is that in theory the iOS XMPP app experience can be improved relatively easily, whereas the fundamental problems around Matrix would be much more difficult to fully resolve.
fundamental problems around Matrix
Sorry, this list just re-hashes the usual talking points and cranks the nerdery to 11. "Reliance on JSON" is just an stupid of an argument as criticizing XMPP for using XML. Saying "the main client uses Electron", presenting it like it is the only alternative (it's not), focusing on a completely secondary platform (pretty much everyone wants to use a messenger on mobile, so why does it matter to have an optimized client on the desktop?) makes this whole thing seem like it was written by a teenager arguing over racing cars.
in theory the iOS XMPP app experience can be improved relatively easily
Then why isn't it?! Why is it that the "recommended" client for iOS looks like something from 2005?
Then why isn't it?! Why is it that the "recommended" client for iOS looks like something from 2005?
That's like asking why this tractor trailer truck still uses the same old boring shipping containers from 40 years ago.
The reference port is the reference port. When FB did xmpp, Google could read their emoji, for instance. It was a brief window before they both diverged from the standard for enshittification, but in that window it all just worked.
That’s like asking why this tractor trailer truck still uses the same old boring shipping containers from 40 years ago.
My point is that most people in 2025 are expecting a lot more from their communications client than what was available in the "good old days" of Gtalk and FB interoperability, and yet most XMPP advocates just bury their heads in the sand and say "it works for me and does everything I need, so there is no reason to add more features".
I'm more concerned with the delays and resource usage, personally.
Then why isn’t it?!
The same reason any open-source project that's community driven isn't as polished as it could be; there's only so many resources and time available to improve things, and much of it is reliant on individuals with the needed skillset taking it upon themselves to improve something that bothers them.
Apparently not enough XMPP developers use iOS for them to focus on polishing up that experience.
I’m more concerned with the delays and resource usage, personally.
As you wish. But you do understand that you are part of a very tiny minority and that it's completely pointless to try to convince people to switch by basing your arguments on the things that align with your values, right?
I'm not using resource usage as an argument for an end-user to switch, the end-user usually doesn't care about technical details. That's more for people hosting servers, who are likely to care somewhat.
The message delays would effect an end-user, and they'll determine if that's too much of an inconvenience or not for themselves. I'm pointing to it as an example of a technical problem that impacts users that seems harder to fix compared to polishing the iOS XMPP app, which was what I initially put forward in my first comment.
If I was to make an argument for why someone should switch away from Matrix, I'd just point to how often encryption fails (I've had continual and significant issues with this personally) and how laggy servers can be. If that isn't effecting someone who isn't technically inclined, then they have no real reason to switch if everything is working for them.
That’s more for people hosting servers, who are likely to care somewhat.
The whole argument at the top of the thread is "it doesn't matter how easy/good/performant it is to run a XMPP server, if the people you'll want to talk to do not have a decent client".
If I was to make an argument for why someone should switch away from Matrix, I’d just point to how often encryption fails.
That has stopped for me since I started using the element X client (mobile) and Fractal (desktop), which leads me to my "counter argument", which is: Matrix has always sucked in different ways, and it sucks for everyone equally, but at the least it is slowly and constantly sucking less than it used to. XMPP is adequate (not great) for some (Android, Linux) and disgraceful for others (iOS), but unlike Matrix there is no visible progress for those on the sucky side of XMPP.
The whole argument at the top of the thread is “it doesn’t matter how easy/good/performant it is to run a XMPP server, if the people you’ll want to talk to do not have a decent client”.
Yes. I don't dispute that an end-user will opt for a more feature-full option regardless of technical problems behind the scenes over a less feature-full option.
My argument is that I think Matrix's fundamental problems that concern the people hosting Matrix servers will eventually, limit its growth, as those problems will be much harder to fix. In contrast; improving an iOS client app is, from a purely technical perspective, much simpler to achieve. I am not making a statement that end-user's will be interested or care about technical details enough to switch their chat platform of choice, unless those details degrade their experience to the point of making them not desire to use it.
I don't use apple products and do not have a good idea of how much progress is being made in that area. It is a legitimate problem that the iOS app is not as polished as other platforms, I do not dispute this. My only point is that it is technically a simpler problem to remedy, were there enough funding or interest by a skilled developer.
If you are not concerned with what I am about the Matrix protocol and its problems, then that's okay. We have differing opinions is all.