I was recently pulled onto Signal by a non-techie who values his privacy. I talked to him about Matrix/Element and he had no idea what that was, but was very happy with Siganl. I must admit, the app is very nice. All I had to do was give it access to my contacts and bam, I am now able to chat with all my contacts.
By comparison, Element is much more like a chat program than a phone messenger. It's good for "I want to connect with that person from GitHub" instead of "messaging the cute girl I met last night" or "messaging my grandpa". And yet, it feels to me like Matrix/Element is the platform less likely to pull something like this. Then again, Keybase seemed that way as well...
> By comparison, Element is much more like a chat program than a phone messenger. It's good for "I want to connect with that person from GitHub"
Element is what messaging should have been from the START: a federated service just like email, where you register an account with your provider of choice, just like email, and start adding/chatting other people after getting to know their address, just like email. So, instead of asking that cute girl her phone number or her email address, you would ask her her element address.
Whatsapp spoiled this approach years ago, so now we are basically screwed because everyone is used to the central approach and it's almost impossible to move away from it. But TODAY's implementation of Element and their shiny clients 12 years ago, would have been a great success just like WhatsApp was (whishful thinking at its finest, I know).
There's also DeltaChat: It looks more or less like WhatsApp, but it uses email as the transport and storage mechanisms, and it is seamlessly encrypted with AutoCrypt. It supports both one-on-one and group chats. It has apps for mobile and desktop.
But I said with today's Element UI/UX. Anyway email was already a standard before corporations took over internet, it would have been really difficult to have a decentralized standard taking over in the 2009 Internet already. Also XMPP was EEE by both Google and Facebook around that time.
I think you hit the nail on the head with "Element is much more like a chat program than a phone messenger". Me and a friend experimented chatting with Element (Riot at the time), and while it certainly "worked", the process of getting everything working was not something I would expect a non-programmer to be able to figure out. We had to finagle different keys across different computers and phones and it was fairly painful. Both of us are software engineers, so at some level we have fun figuring this stuff out, but I cannot see a universe where Element catches with the general public unless the process is as quick and painless as Signal.
I feel like Element works better as a competitor to Slack or IRC than as a competitor to Signal or Whatsapp.
> I feel like Element works better as a competitor to Slack or IRC than as a competitor to Signal or Whatsapp.
To me it's a competitor to Keybase. "I want to send my co-worker/client an API key that I don't want exposed to the public" is about the only use for Keybase I've had. I have like 5 contacts on there for this reason. Slack/IRC is much more usable for getting shit done, but not being E2E I wouldn't send anything sensitive over them. Element is currently a "this is a mildly nicer experience over PGP + Email/Slack.
I know very little about the intricacies of cryptography, but part of me wonders if there's some way of doing a federated "key synchronization" service similar to keybase.
So Keybase is just a UI for PGP/GPG (well that was what it was before it became a Borg). The problem with GPG:
1. You need to keep your private key very private, which is incompatible with the idea that you might have several devices you normally use. GPG itself does not provide you with a mechanism to sync your private keys between devices because this is a super insecure thing to do without some serious work.
2. GPG requires that you and another person verify each others' public keys out of band. I need to meet you in a parking lot to validate your key fingerprint while you validate mine.
3. GPG's web of trust relies on attaching public keys to real world identities. You are asked to validate government documents when verifying public keys. That's incompatible with how a lot of us want to work. Note that this isn't a built-in requirement, but GPG itself provides no guidance on how to validate user123 on GitHub, just User Onetwothree Jr in real life.
4. GPG's UI is almost as arcane as tar :)
Keybase solved this by:
1. Providing a secure way to manage private keys across devices.
2. Outsourcing proof of identity to other providers. Its use case is validating the identity of user123 on GitHub, which happens to also work fairly well for CelebrityName on Twitter, or FriendName on Facebook.
3. See #2: social proof means you can attach that proof to any kind of identity.
4. GUI + nice TUI works better.
Where Keybase fell short was that a non-techie will not understand much about "social proof" and the only kind of social proof they have access to is limited to Twitter, Facebook, and Instagram.
Signal's solution to this was simpler: you have a QR code/set of numbers that represent your fingerprint right in the app. You show me yours, I'll show you mine. We get connected by phone number or email. That's it. If Signal was built on a federated platform it'd be perfect and nothing about it from what I understand prevents that.
That sounds kinda similar to the problem Matrix solved with cross-signing, how when you login to a new device and verify it with one of your already logged-in devices, it can request your old message keys E2EE so you get all your history.
Maybe a similar thing could be built on top of it?
It has improved a lot. I once wanted to switch with another tech-savvy friend 2 or 3 years ago and the experience was abhorrent.
Nowadays I use it mostly like a IRC client and it improves constantly.
However the comparison between this and signal falls flat due to the metadata that needs to be stored on matrix servers due to its federated setup.
They are less likely to do this kind of secretive development, but they could go that direction. They have considered cryptocurrency in the past, see https://matrix.org/blog/2017/08/22/thoughts-on-cryptocurrenc.... They are open, but still driven by a single company which could change direction at any time.
They also surprised their community multiple times with renames of their app and weird redesigns (remember the horizontally-scrollable unordered bubbles for room selection?)
Which part? The part of also integrating SMS functionality? The part where I can message my other contacts who aren't using Matrix via SMS and finding them by phone number? Having a good marketing strategy?
Sorry, I should have quoted. They are “unable to pull something like this” due to being a federated protocol. If they try to add crypto to their app, another app can be used to communicate to the same people in the same way.
Would you say this is a weakness of element or matrix itself? In principle you could made a clone of signal, WhatsApp, telegram etc. using mobile APIs right?
I think it's 100% the client. But this is the problem with a federated system like this: it increases your marketing surface without providing apparent value to the consumers. Consumers don't want choice, they want the one product that will do exactly what they need it to do. When I am presented with "choose your client from this list of 5-15" my eyes glaze over. I just want to try the thing. That isn't to say that there shouldn't be choices, there absolutely should. But the problem is that there needs to be a very easy short and gentle on ramp for new users.
Element is none of those things. It's name is so forgettable and so generic that people often don't even know whether it's an app, a library, a website, etc. The mobile app is yet another chat app with nobody on it until I do the legwork of pulling them in. It's just not usable on day one after I already spent the time to figure out which app I need. In the meantime, Signal can become your default messenger on Android within a few minutes and do everything you used to be able to do but more and better.
By comparison, Element is much more like a chat program than a phone messenger. It's good for "I want to connect with that person from GitHub" instead of "messaging the cute girl I met last night" or "messaging my grandpa". And yet, it feels to me like Matrix/Element is the platform less likely to pull something like this. Then again, Keybase seemed that way as well...