It would be nice if the core functionality of the iOS client[1] and accessibility of the ecosystem were fixed too, or at least if some visible movement was seen on those. The basic ability of two mutually encrypted users to send each other messages has been broken for months with very little responsiveness from the maintainers.
One of the key issues I have found in general with the matrix ecosystem is that it is highly inaccessible to the people who need it most – non-technical but tech-literate end users. For instance its SDKs and APIs and protocol are all very well (even beautifully) documented for the most part, but client functionality, less so – there is no documentation on the website or anywhere I could find in the repos discussing e.g. how to properly set up cross signing, or even going through the basic steps of how to verify another user’s device (not to mention why one would want to bother, or what E2EE is in layman’s terms)!
Coupling that with removing “escape hatch” functionality like the ability to manually verify someone from the mobile clients, and it becomes exceedingly difficult for non-technical users – even for more technical users like me – to understand when something isn’t working because of improper setup on the part of the user, or because the client has been broken for months as the iOS client has been.
The end result is that the entire ecosystem goes from being a genuinely useful tool for activists, journalists, and vulnerable demographics to being basically a toy whose benefits can only be properly enjoyed by more technical users for whom none of this is a matter of life and death. Blog posts like this, while interesting for someone like me who has the background to understand them, only reinforce this: very few of the blog posts on matrix.org are comprehensible to end users who aren’t also engineers.
arathornMatrix.org project lead
edited
3 years ago
|
link
So there’s a big difference between Matrix the protocol, Synapse the server implementation (subject of the OP here) and Element the client (subject of the comment here). The Matrix.org website and Synapse and the OP are unashamedly focused on developers and sysadmins.
Separately, totally agreed that matrix clients like Element need to do better at usability - the intention is of course to be as easy as WhatsApp / Telegram / Slack etc to use, and there should have some major changes in the next weeks to show visible progress there. Meanwhile, the bug you linked is a plain old bug which we’re hunting down (thank you for providing the logs).
I appreciate the reply. I’m aware of the differences between the protocol and the server and client implementations. That said, I’m genuinely excited to hear about those updates, and to know that you agree there’s a usability problem. Thanks for working on that bug.
Synapse was dog-slow with rooms with a large number of people for a long time, which made the experience of using Matrix frustrating and unreliable. Fixing that is as important for usability by anyone, including nontechnical people, as fixing client-side encryption.
The end result is that the entire ecosystem goes from being a genuinely useful tool for activists, journalists, and vulnerable demographics to being basically a toy whose benefits can only be properly enjoyed by more technical users for whom none of this is a matter of life and death.
People who understand encryption well enough to deal with Matrix client encryption usability issues can just as easily be activists, journalists, and people who would be harmed by having their private communications leaked, as people who don’t. These are orthogonal, not disjoint, categories.
Thanks for the point about performance being relevant from an accessibility perspective
With regard to your second point, the set theoretic language obscures more than it helps here since we’re talking about accessibility. These sets are fuzzy: clearly activists, journalists, and others who aren’t software engineers, security researchers, or otherwise have technical (or tech-adjacent) day jobs can and do deal with Matrix client encryption. My claim is that Element is, at the time of writing this, orders of magnitude more challenging to learn, use, and trust if you aren’t in the latter category. For some, my observed experience is that the usability concerns make it completely non-viable.
The fact that some e.g. journalists can persevere and make it work for them is not relevant to my point – that is similar to saying “even though Acme Cafe isn’t wheelchair accessible, it’s still accessible to other people”.
It would be nice if the core functionality of the iOS client[1] and accessibility of the ecosystem were fixed too, or at least if some visible movement was seen on those. The basic ability of two mutually encrypted users to send each other messages has been broken for months with very little responsiveness from the maintainers.
One of the key issues I have found in general with the matrix ecosystem is that it is highly inaccessible to the people who need it most – non-technical but tech-literate end users. For instance its SDKs and APIs and protocol are all very well (even beautifully) documented for the most part, but client functionality, less so – there is no documentation on the website or anywhere I could find in the repos discussing e.g. how to properly set up cross signing, or even going through the basic steps of how to verify another user’s device (not to mention why one would want to bother, or what E2EE is in layman’s terms)!
Coupling that with removing “escape hatch” functionality like the ability to manually verify someone from the mobile clients, and it becomes exceedingly difficult for non-technical users – even for more technical users like me – to understand when something isn’t working because of improper setup on the part of the user, or because the client has been broken for months as the iOS client has been.
The end result is that the entire ecosystem goes from being a genuinely useful tool for activists, journalists, and vulnerable demographics to being basically a toy whose benefits can only be properly enjoyed by more technical users for whom none of this is a matter of life and death. Blog posts like this, while interesting for someone like me who has the background to understand them, only reinforce this: very few of the blog posts on matrix.org are comprehensible to end users who aren’t also engineers.
[1] https://github.com/vector-im/element-ios/issues/3762
So there’s a big difference between Matrix the protocol, Synapse the server implementation (subject of the OP here) and Element the client (subject of the comment here). The Matrix.org website and Synapse and the OP are unashamedly focused on developers and sysadmins.
Separately, totally agreed that matrix clients like Element need to do better at usability - the intention is of course to be as easy as WhatsApp / Telegram / Slack etc to use, and there should have some major changes in the next weeks to show visible progress there. Meanwhile, the bug you linked is a plain old bug which we’re hunting down (thank you for providing the logs).
I appreciate the reply. I’m aware of the differences between the protocol and the server and client implementations. That said, I’m genuinely excited to hear about those updates, and to know that you agree there’s a usability problem. Thanks for working on that bug.
Synapse was dog-slow with rooms with a large number of people for a long time, which made the experience of using Matrix frustrating and unreliable. Fixing that is as important for usability by anyone, including nontechnical people, as fixing client-side encryption.
People who understand encryption well enough to deal with Matrix client encryption usability issues can just as easily be activists, journalists, and people who would be harmed by having their private communications leaked, as people who don’t. These are orthogonal, not disjoint, categories.
Thanks for the point about performance being relevant from an accessibility perspective
With regard to your second point, the set theoretic language obscures more than it helps here since we’re talking about accessibility. These sets are fuzzy: clearly activists, journalists, and others who aren’t software engineers, security researchers, or otherwise have technical (or tech-adjacent) day jobs can and do deal with Matrix client encryption. My claim is that Element is, at the time of writing this, orders of magnitude more challenging to learn, use, and trust if you aren’t in the latter category. For some, my observed experience is that the usability concerns make it completely non-viable.
The fact that some e.g. journalists can persevere and make it work for them is not relevant to my point – that is similar to saying “even though Acme Cafe isn’t wheelchair accessible, it’s still accessible to other people”.