I'm 🤏🤏🤏 this fucking close to consider matrix unusable and ditching it
As a user this is completely unacceptable and I genuinely don't give a single, tiniest amount of shit why it's happening. I just don't care. This has no right to happen if this piece of garbage software/protocol wants to be considered usable in the real world.
What happened: Talking to someone
Account 1: issues with images. They can't see the files I sent. Issue is not replicable when I try to debug it with my instances admin
Account 2: this shit
This is not a bad UX
This is catastrophically bad UX beyond repair
Once a person tries to touch Matrix and experiences this motherfucking joke of an experience, they will very rightfully not want to touch it ever again
I only used it for so long because I'm dedicated and like self-hostable communication and EVEN I am at my breaking point with it
@rail_ The frustrating part is that almost the entire problem lies with Element's stuff specifically, but it's so prevalent and woven into everything that it pretty much defines the image and experience for all of Matrix.
(And also I've warned the Element folks about exactly this outcome *so many* times and they have not acted on it)
@anthropy @rail_ All of that ultimately leads back to Element, unfortunately.
I know a lot of the folks working on alternative implementations, and almost without exception they start out with great intentions and then somewhere along the way realize that the spec is incomplete / Synapse is unreliable / the documentation about E2EE wrong / there is a client interop issue / etc.
Even the non-Element clients are constantly held back by the mess on a governance level, making it way more difficult than necessary to implement a good client, and nearly impossible to implement a reliable homeserver. The result is that people end up burning their time and energy on accounting for all these issues so that it works at all, instead of on building nice software for Matrix.
There's just no fast solution for this that I can see. The possible paths to a solution that I know of are either a) serious changes in governance (ongoing but slow), or b) just spending years on mapping out all these issues and working around them to build a stable base. Once either of those points are achieved, building good software is possible.