1y ago

(15th of Dec) Element Discontinuing Hosted Matrix for Consumers

You're viewing a single thread.

14 comments

While unfortunate for consumers I get it. Usually consumer customers use more support than business ones while paying significantly less and element still has to make enough money to keep going.
- Yeah, I hope Element manages to become sustainable. It feels like they are running out of funding from what they did in the last month.
  
  I don't see the reason we need a venture capital funded bloated protocol anyways. Just switch to XMPP. It's much more lightweight and it's the internet standard for instant messaging.
  
  Let's clear something up:
  Matrix is a free and open protocol.
  Element is one of several free and open source Matrix clients.
  Vector Limited is a company funding the development of the Element client, that also sells a Matrix homeserver.
  There are several other free and open source Matrix homeservers.
  You DO NOT NEED to pay a dime to Vector Limited for a Matrix homeserver, or use the Element client.
  Meanwhile, XMPP is a "built by committee" nightmare, with a committee that didn't commit (pun intended) to basic features, leaving file transfer, audio, or video calls, to be defined as "protocol extensions" (aka: incompatible) by each client.
  TL;DR: use Matrix, with the Element client or any other, leave Vector Ltd. for the businesses.
  
  “protocol extensions” (aka: incompatible)
  Reality shows that implementations can very well implement the same extensions. If you don't use extremely outdated clients you will find they do have compatible file transfer and A/V calls. ActivityPub works the same way.
  Meanwhile Matrix Ltd. cooks up a completely new, incompatible protocol instead of building upon existing internet standards.
  
  There is no "Matrix Ltd.", the Matrix protocol is being worked on by the "Matrix Foundation", a UK CIC (kind of NGO), with adoption by French an German governments.
  XMPP clients "appear" to have compatible file transfer and A/V calls... until you try using them and find out they sometimes lose bytes from one client to another, but not the other way around, sometimes calls only work one way, and so on. That's the effect of not having a minimum common ground defined in the protocol spec.
  
  The NGO is a decoy organization with exactly the same people (minus one) as the VC funded startup. Go look at the "core spec team" and find out which organization they belong to.
  Your information on XMPP seems to be quite outdated. File transfer in XMPP is now mostly done by uploading the file via HTTP and sending the URL. Audio calls are done using WebRTC and work two ways.
  
  I wish XMPP had stuck around. I used to run a Prosody server and it worked well enough but I think the E2E keys would occasionally need to be fixed. I used Conversations on Android as a client at the time. The things that makes me hesitate to dedicate too much effort to Matrix are:
  the supposed funding issues they're having (which is part of why I paid for hosting)
  the FOSS' communities seeming tendency to keep jumping messaging platforms and so there's never a chance for one to gain critical mass
  how buggy the web client and Element iOS client have been.
  When I stopped running an XMPP server I switched the only other user over to Signal and we've stuck there since. With how buggy the Element iOS client, Fluffy Chat and web client have been for me (app crashes when joining rooms, rooms don't exist when they in fact do), I don't want to risk an upset by trying to push people there since Signal is good enough. And these are all issues that exist when the company who makes Matrix (plus contributors of course) are the ones running the server.
  At this point I'm just inclined to grab the export they provide and switch to matrix.org for the 1 or 2 rooms I care to have a presence in.
- But on the other side, existing customers don't use much support, once everything is set and working, it's all profit for them.
  Kicking out customers equals to giving up money. I don't think many of them would start paying double (from $3 per user to $5 per user) or more (you gonna pay for 20 users even if you have 10) instead of just renting a VPS and do it for free
  
  I can assure you existing customers(non business) do still use a disproportionate amount of support. Things like feature requests, lost passwords, new user on boarding, any time anything doesn't work as the user expects, etc
  As for the finances here sure it gives up some income but will also free up a lot of time for someone who is currently doing support so they could focus on other things
  
  Surely, password stuff is automated no?
  
  I disagree, they could reply with automated canned responses if don't have resources for support
  "Suggest new feature" => "thanks we will consider" => trash it
  "Help I'm locked out" => "use this self reset tool or cancel your account"
  "I need to add a new user" => "sure, read the documentation"
  There are many other paid services without any kind of support in the entry level plan, could have done that. For example the $24/year plan for photoprism doesn't come with support

14 comments