191–199 of 199 posts
So, even if you believe Matrix is the future and Signal/WhatsApp/Telegram have a drastic flaw, we couldn't get here without them.
Earlier quoted context omitted.
> _why_ communication has moved away from email Has it, though? I don't think we are any step closer now to dropping email than we were 25 years ago. > or why XMPP/IRC never really caught on despite being vastly older than these app competitors Could be the lack of marketing (because there isn't a corporation or venture capital money behind benefiting from exponential user base growth) and critical mass/peer-pressure…
> Has it, though? I don't think we are any step closer now to dropping email than we were 25 years ago. I think OP meant that there's been a significant decrease in usage. In my case I can definitely relate to this for personal communications, I no longer send emails to personal acquaintances, I mostly use email for professional purposes, so I think OP's point stands. > Could be the lack of marketing (because there i…
I think we can agree that they have different strengths, I haven't seen much change in my usage of email for chatting with personal acquaintances because email has never been going strong there, unlike Instant Messaging…
> but a killer feature is way more efficient, especially in the long run
…if one thing, popular IM of today pack less features, not more, than those they replaced. Remember the golden age of MSN messenger? You had whiteboards, tictactoe/chess/… games with friends, screen sharing, fancy fonts & colors, "now listening to:" , wizz, … Ironically, XMPP inherited all those features out of necessity to be compatible with all those protocols (and I don't think it matters the slightest).
XMPP does have quite nice and unique features of its own, if check-out the other response :)
> WA offered users a way to text using the same phone numbers for free in countries where you had to pay
just like every other messenger then and now… (I'm not downplaying their coming ahead of this fierce competition, they must have done something right indeed, but "IM as a cheap alternatives to SMS" was popular a decade before WA)
> WA also uses the Signal Protocol, which I assume is not directly compatible with OMEMO
"Signal Protocol" is a mashup of encryption algorithms, signaling and session negotiation, at its very core there is the double-ratchet algorithm (for forward secrecy) and prekeys (for offline delivery), OMEMO is just that, implemented over XMPP primitives (with prekeys served over PubSub/PEP).
> My point is that it's not as simple as setting COMPATIBLE=TRUE to make every network compatible with each other.
Indeed, it would take some careful considerations, but eh, same can be said about TLS. At a first glance, I don't see anything radically difficult about it, the main point would probably be to offer a generic endpoint to serve prekeys and there's nothing revolutionary or difficult about that.
> I doubt Facebook deviated from the base protocol just for the sake of it
They did what every big business does: they optimized for their specific needs. When they decided that compatibility wasn't something they could monetize, they shut the gateway down and stopped caring about it. That tells nothing about the protocol, its capabilities, or of a presumed inability to extend it. For that matter, the X in XMPP stands for "eXtensible" protocol, and it's been quite good at keeping up for about 22 years, now :)
Earlier quoted context omitted.
Sure! The first "practical"* difference between Matrix and XMPP is that Matrix' chatrooms are "necessarily decentralized" (whereas in XMPP that's optional [1]). How it works is that users involved in a given room have their servers synchronize the room's state with each other (room state is messages, presence status, read receipts, "is composing" status, …) so when one server goes down/netsplits, messages, join/leave…
Thanks, I'll have to dig into the protocol, maybe do some sniffing on my homeserver :) I assume not every node gets the metadata for all the rooms in the world, just the ones its users participate in? Anyway looking forward to investigating this.
Earlier quoted context omitted.
I was designing something similar to Matrix at one point several years back. The different names confused even myself when I was dabbling into Matrix. Matrix is cool, there's no doubt. The packaging is terrible, IMO. They really need a product manager.
Do you have any links to share?
Earlier quoted context omitted.
Here's an idea, for anyone reading this to run with. Sell account packs, say 50 accounts for 10 bucks. Or maybe 10 accounts for 5 bucks. Or maybe 20 accounts for $30 a year? Adjust it for both sustainability with your tech model, I dunno about that. But the idea is to minimize the number of payments required per account and per network, and allow one person to pay for their entire group.
Doesn't Element do this already? https://element.io/plans-and-pricing/pro
Earlier quoted context omitted.
> Sure but that does not in any way take away from the point that siloed messengers are a bad idea. Well, too bad! Usability trumps idealism when network effects are at play. > If the XMPP/Matrix people wanted to they could set up a trusted third party server to link identities to phone numbers just like Signal does Usability is so bad that you're defending Matrix and don't even know this is already available, just h…
Have you tried other clients than Element yet? For phone, you could e.g. try FluffyChat or Ditto, and for Desktop you could try e.g. nheko or mirage
Ditto has a dead website and it's not on the Play store, there's a link on its repo but it's dead. APKs are not viable for mainstream use.
And I just tried onboarding someone on FluffyChat, and as of right now, it's non-viable as well. During the process I have to go back and forth between web browser and application which completely went over my mother's head despite the textual instructions, and when it's set-up it defaults to Matrix usernames instead of doing automatic contact resolution using an identity server.
On the other hand, the UI is pretty nice and fits in well with native apps, if the app had a more serious, bland branding and a tamer color palette it would be up to task UI-wise.
I really sympathise with this point of view, but when it comes to messaging with family and friends, I just don't really care. Matrix seems to want to be an IRC replacement rather than a one-to-one / small-group chat app. That's fine, but often not what I want. Teaching my mother in law to use Signal: Install. Accept permissions. Choose a PIN. I can send you messages. You can see your friends who already use Signal.…
Moments of recognition reading this. A couple of weeks ago at virtual FOSDEM I was trying to work out how to join the FOSDEM chatrooms using the Android app formerly known as Riot.im (now Element.io). And I couldn't work it out - I gave up completely. I'm an actual programmer and heavy IRC user!
Earlier quoted context omitted.
XMPP lost popularity long before the public became (unhealthily) obsessed with encryption.
Care to explain? I can't think of a single example where encrypting a private message passed over a public network would be unhealthy.
Telegram's far better UX than Whatsapp is possible because they do not have e2ee on anything but Secret Chats (and I've never seen anyone on Telegram using them). See, for most users the promise of security is better than actual security and all attached UX problems.