7 ms·

How about XMPP support? I mean, you are talking about Matrix being the holy grail of chat and at the same time you do not support the IETF standard for instant

How about XMPP support? I mean, you are talking about Matrix being the holy grail of chat and at the same time you do not support the IETF standard for instant messaging (which is also federated, supports E2E encryption, can use bridges to other networks and has several open source implementations)... In general, my biggest issue with the Matrix community is that they chose to build something new, instead of fixing something existing. Granted, at the time when Matrix started, XMPP wasn't fit for the mobile revolution. But instead of improving the XMPP standard (which other people did afterward), the Matrix people decided to build something new from the ground up. They took a few different design decisions, but in my opinion, nothing that would justify building a competing solution and splitting the already thin developer community. Now we have too solutions, both failing to find significant adoption. I understand that the matrix people probably built their eco system as a hobby, so who am I to criticize them. I just feel so depressed, seeing so much work being done on a very important subject, not fulfilling its potential due to missing focus.

acct776 6y ago
Was XMPP improved at that time? Or a mess of nonstandard addons/plugins that you had to match up, like mods in a multiplayer videogame?
- Semaphor 6y ago
  Close. It’s a mess of standard addons/plugins that you have to match up, like mods in a multiplayer videogame. Sadly for me, it’s also the only modern messenger that has decent desktop clients that don’t look like discord. But I guess that this will never change as I seem to be pretty lonely with my want of those.
  - dijit 6y ago
    The solution was nearly always to create golden packages which could be summarised. XMPP v1 server could be just XMPP itself. A v2 client/server includes packages for OMEMO, Websocket, Resumeable connections A v3 client/server has packages for whatever else becomes the de jour. The idea that you can just mix and match without ever standardising on a set of capabilities was a large part of what killed XMPP and was entirely solvable.
    - MattJ100 6y ago
      This is what happens. Compliance suites are published annually: https://xmpp.org/about/compliance-suites.html https://xmpp.org/about/compliance-suites.html
      - Semaphor 6y ago
        This is either barely getting used or not used enough (or maybe too hidden). Because I had only seen single XEP lists for both clients and servers I looked at.
- pseudalopex 6y ago
  They said XMPP improved afterward. The point is a combined effort would be in better shape than XMPP or Matrix now. The problem with XMPP was every client implemented different parts of the standard. The problem with Matrix is every client implements different parts of the standard.[1] [1] https://matrix.org/clients-matrix https://matrix.org/clients-matrix
  - acct776 6y ago
    Not being able to share Rick and Morty stickers or have a group call is fine to miss out on. Having to choose from multiple encryption addons to get started even chatting, that's not a comparable experience. I miss XMPP, but it's time to let go.
kitkat_new 6y ago
> They took a few different design decisions, but in my opinion, nothing that would justify building a competing solution and splitting the already thin developer community. I'd argue that the design decisions are key to the success of Matrix. However I do not think that there must be any splitting. Checkout the XMPP Bridge that matrix.org hosts. > I just feel so depressed, seeing so much work being done on a very important subject, not fulfilling its potential due to missing focus. As long as there are people like Eric, there is hope :) I think we have more focus than ever before, even within the XMPP community :)
- pseudalopex 6y ago
  The XMPP bridge means the user communities don't have to be split. The developer communities are.
- upofadown 6y ago
  >Checkout the XMPP Bridge that matrix.org hosts. Is there a description of how to generate an address on the other network? This sounds awesome. A working bridge gets a Matrix user access to the tens of thousands of federated XMPP servers.
  - kitkat_new 6y ago
    Yes, there is: https://github.com/matrix-org/matrix-bifrost/wiki/Address-syntax https://github.com/matrix-org/matrix-bifrost/wiki/Address-sy...
Arathorn 6y ago
Speaking as a Matrix person: we're quite happy with our adoption, which is accelerating exponentially, and we didn't build Matrix as a hobby: it's been the team's fulltime day job since 2014. Before that we used XMPP (ejabberd + spark + XMPP.framework etc) and eventually decided to build a totally different architecture: one focused on syncing conversation history, rather than sending instant messages. I don't think it dilutes or splits the thin developer community: instead it's increased interest in open comms enormously (as well as helping spur the XMPP community into improving their stuff). Just as Linux didn't somehow destroy BSD.
- selfhoster11 6y ago
  I can only thank you for your work. Open messaging ecosystems are much better off for it.
- arendtio 6y ago
  Well, even if I prefer XMPP over Matrix and still disagree over the developer ressources, I wish you the best of luck, because both solutions, being federated, are inherently better than the competition by design.
jiofih 6y ago
Because XMPP sucks. It had 20 years to succeed, and failed. Maybe it’s time for you to move on? Last company I worked at that used it internally had thousands of employees in IT and still considered maintaining the jabber server not worth it - a million XEP extensions cannot bring it to the same level as a modern chat client. On every matrix post the xmpp crowd comes out of the woods, but what have you got to show? What’s the best cross-platform or web client for it right now? They all look like half-baked ports of long abandoned QT apps made for Linux and still don’t support syncing/multi-presence well.
- pseudalopex 6y ago
  Did the company switch to Matrix? XMPP was pretty successful in the 2000s. More than Matrix so far. There were missed opportunities but the main reason it declined was Facebook and Google decided closed was more profitable than open. Why exclude native clients?
  - SulfurHexaFluri 6y ago
    I'm not even convinced they decided it was more profitable. The decision probably went something like this: Manager: Our users want this feature Developer: We can't add that because it will cause incompatibility with XMPP clients Manager: This is the 50th feature that has been blocked by XMPP and hardly any of our users use it. Time to cut off support for it
    - mszcz 6y ago
      That and they probably threw hoards of cash and engineering at the problem.
    - xorcist 6y ago
      It's probably much simpler than that: "It has come to my attention that clients which we do not control connect to our service. This severely limits the monetization opportunities available to us. Please fix before it is caught in a due diligence."
  - 6y ago

acct776 6y ago

Was XMPP improved at that time? Or a mess of nonstandard addons/plugins that you had to match up, like mods in a multiplayer videogame?

Semaphor 6y ago
Close. It’s a mess of standard addons/plugins that you have to match up, like mods in a multiplayer videogame. Sadly for me, it’s also the only modern messenger that has decent desktop clients that don’t look like discord. But I guess that this will never change as I seem to be pretty lonely with my want of those.
- dijit 6y ago
  The solution was nearly always to create golden packages which could be summarised. XMPP v1 server could be just XMPP itself. A v2 client/server includes packages for OMEMO, Websocket, Resumeable connections A v3 client/server has packages for whatever else becomes the de jour. The idea that you can just mix and match without ever standardising on a set of capabilities was a large part of what killed XMPP and was entirely solvable.
  - MattJ100 6y ago
    This is what happens. Compliance suites are published annually: https://xmpp.org/about/compliance-suites.html https://xmpp.org/about/compliance-suites.html
    - Semaphor 6y ago
      This is either barely getting used or not used enough (or maybe too hidden). Because I had only seen single XEP lists for both clients and servers I looked at.
pseudalopex 6y ago
They said XMPP improved afterward. The point is a combined effort would be in better shape than XMPP or Matrix now. The problem with XMPP was every client implemented different parts of the standard. The problem with Matrix is every client implements different parts of the standard.[1] [1] https://matrix.org/clients-matrix https://matrix.org/clients-matrix
- acct776 6y ago
  Not being able to share Rick and Morty stickers or have a group call is fine to miss out on. Having to choose from multiple encryption addons to get started even chatting, that's not a comparable experience. I miss XMPP, but it's time to let go.

kitkat_new 6y ago

> They took a few different design decisions, but in my opinion, nothing that would justify building a competing solution and splitting the already thin developer community. I'd argue that the design decisions are key to the success of Matrix. However I do not think that there must be any splitting. Checkout the XMPP Bridge that matrix.org hosts. > I just feel so depressed, seeing so much work being done on a very important subject, not fulfilling its potential due to missing focus. As long as there are people like Eric, there is hope :) I think we have more focus than ever before, even within the XMPP community :)

pseudalopex 6y ago
The XMPP bridge means the user communities don't have to be split. The developer communities are.
upofadown 6y ago
>Checkout the XMPP Bridge that matrix.org hosts. Is there a description of how to generate an address on the other network? This sounds awesome. A working bridge gets a Matrix user access to the tens of thousands of federated XMPP servers.
- kitkat_new 6y ago
  Yes, there is: https://github.com/matrix-org/matrix-bifrost/wiki/Address-syntax https://github.com/matrix-org/matrix-bifrost/wiki/Address-sy...

Arathorn 6y ago

Speaking as a Matrix person: we're quite happy with our adoption, which is accelerating exponentially, and we didn't build Matrix as a hobby: it's been the team's fulltime day job since 2014. Before that we used XMPP (ejabberd + spark + XMPP.framework etc) and eventually decided to build a totally different architecture: one focused on syncing conversation history, rather than sending instant messages. I don't think it dilutes or splits the thin developer community: instead it's increased interest in open comms enormously (as well as helping spur the XMPP community into improving their stuff). Just as Linux didn't somehow destroy BSD.

selfhoster11 6y ago
I can only thank you for your work. Open messaging ecosystems are much better off for it.
arendtio 6y ago
Well, even if I prefer XMPP over Matrix and still disagree over the developer ressources, I wish you the best of luck, because both solutions, being federated, are inherently better than the competition by design.

jiofih 6y ago

Because XMPP sucks. It had 20 years to succeed, and failed. Maybe it’s time for you to move on? Last company I worked at that used it internally had thousands of employees in IT and still considered maintaining the jabber server not worth it - a million XEP extensions cannot bring it to the same level as a modern chat client. On every matrix post the xmpp crowd comes out of the woods, but what have you got to show? What’s the best cross-platform or web client for it right now? They all look like half-baked ports of long abandoned QT apps made for Linux and still don’t support syncing/multi-presence well.

pseudalopex 6y ago
Did the company switch to Matrix? XMPP was pretty successful in the 2000s. More than Matrix so far. There were missed opportunities but the main reason it declined was Facebook and Google decided closed was more profitable than open. Why exclude native clients?
- SulfurHexaFluri 6y ago
  I'm not even convinced they decided it was more profitable. The decision probably went something like this: Manager: Our users want this feature Developer: We can't add that because it will cause incompatibility with XMPP clients Manager: This is the 50th feature that has been blocked by XMPP and hardly any of our users use it. Time to cut off support for it
  - mszcz 6y ago
    That and they probably threw hoards of cash and engineering at the problem.
  - xorcist 6y ago
    It's probably much simpler than that: "It has come to my attention that clients which we do not control connect to our service. This severely limits the monetization opportunities available to us. Please fix before it is caught in a due diligence."
- 6y ago