this post was submitted on 21 Sep 2023
250 points (100.0% liked)
Technology
37699 readers
281 users here now
A nice place to discuss rumors, happenings, innovations, and challenges in the technology sphere. We also welcome discussions on the intersections of technology and society. If it’s technological news or discussion of technology, it probably belongs here.
Remember the overriding ethos on Beehaw: Be(e) Nice. Each user you encounter here is a person, and should be treated with kindness (even if they’re wrong, or use a Linux distro you don’t like). Personal attacks will not be tolerated.
Subcommunities on Beehaw:
This community's icon was made by Aaron Schneider, under the CC-BY-NC-SA 4.0 license.
founded 2 years ago
MODERATORS
you are viewing a single comment's thread
view the rest of the comments
view the rest of the comments
"Sliding sync" is Matrix's own admission that the protocol is too complex and taxing on clients to be practical, and shifts the burden further onto already overwhelmed servers for what's essentially bouncers marketed as new tech. And it's still a mess.
I know of no major messaging service where the client wants to download everything
I can’t wait to see what you’ll develop in response!
I won't need to develop anything in response, because an open-standard (IETF) protocol for federated instant communications already existed long before Matrix, and as far as I can tell, from my experience of having administered XMPP and Matrix servers for hundred of users, nothing about Matrix, its design and its implementations makes it more desirable, more reliable, more resilient or more "future proof" than what XMPP came-up with a decade earlier.
And I am aware that I sound like an old man yelling at clouds, I take comfort in the fact that more and more technically-versed people who look behind the marketing and buzz get to see what I know from experience: https://telegra.ph/why-not-matrix-08-07
I think most of the criticism on Telegraph regarding how Matrix handles rooms and events are addressed by the work behind linearized matrix: https://www.qwant.com/?l=en&q=linearized+matrix+messaging&t=web
Since its inception, Matrix has always been "months away" from cracking this very problem, I won't hold my breath for this one, not after 10 years of kicking the same can down the road.
Is their shift key broken?
Furthermore, this blog post has outdated information and many of their problems with Matrix are fundamental for federated protocols. Good luck removing an email sent to another server, for example. JSON form is very well defined.
I can agree with the problem of DAG complexity building up, sure, but that is a tradeoff.
Could be the result of repetitive strain injury, I'm not judging
Yup, I don't think that's the main argument, more a warning like "well if you normal folks were expecting anything else, that's how it is and can't be changed"
Though that's flawed/incomplete, and an essential reason for incompatibility between implementations
That's my main issue, really. Basically, there's no end in sight to the problem of inconsistent messages that can take minutes/hours/days to be delivered: Matrix just isn't dependable for instant communications, by design. This rules it out of a ton of very significant and practical real world use cases and expectations. That's not a system you can think typical modern users (used to their messages being either delivered instantly or not at all) to trust and to like once faced with these issues, not just once, but repeatedly. It is us, the world, being told again "Blockchain will solve all problems of centralization", with a crowd of enthusiasts blinded by the hype embarking (taking hostage) their less savvy peers while the people being the steering wheel have no idea how to bring the car home. And worse, those people behind the steering wheel being in denial (or malicious, or incompetent), will keep telling you to just trust them, and that the problem is solved already while the evidence points in the other direction.
Admitting problems and improving/replacing your protocol is good, you make it sound like a bad thing. I mean you could argue that they should have started with this, but imo better late than never. From what I've seen this will take load off of the client AND the server, because both don't have to sync thousands and thousands of events anymore. It basically looks like an indexing/caching layer between client and server, which is standard practice to make things go faster, especially for thin clients.
The only bad thing about this is that we've been at it for 10 years. If you've been following Matrix long enough, you've witnessed "the next big thing that will solve all problems" being promised every year. Matrix funding relies on hype, and I'm somewhat ok with that, so long as users and hosts are not taken hostage of empty promises. My first hand experience of Matrix X is that we are still far from what's being advertised.
edit: adding a missing word "thing"
I really don't get this attitude. It's not like global decentralized instant messaging with all the usability, bells and whistles of centralized services is an easy problem to solve. And no one is selling anything, not to regular users at least. If you thought that this would be a straight forward path to a finished product then idk what to tell you, that's not how this works.
Yup, absolutely, and being in this space myself as an enthusiast, that's an interesting problem to see being worked on and having significant brain power allocated to, though that doesn't remove anything from the fact that
..is precisely what Matrix developers are advertising Matrix to be, and actively marketing it to be. You can go on hackernews right now and observe Arathorn telling everyone that everything is fine and solved now, even when shown evidence that it is not, like he has done since the beginning.
I believe people should know what they are engaging with.
Idk what stuff you're reading, but every dev talk I've seen includes many acknowledgments of the shortcomings that still exist and the difficulty of the underlying problems. I never had the impression that they're trying to sell a silver bullet that'll fix everything once and for all. It's mostly just incremental changes here and there that fix or improve certain parts of the system, and with that a steady progress towards the goal.
I wouldn't give too much on the speculations and opinions of any one person, even if he's the Matrix project lead. Probably especially the project lead, because part of his job is being optimistic about changes so they actually happen. But this is still mostly uncharted territory, and all anyone can do is make best effort attempts to improve things bit by bit. And from what I've seen he also openly talks about issues and the limits of coming changes, so perhaps you just read too much into his more optimistic posts and comments?
Personally I'm just excited for new developments, but also aware that any change has to prove itself in the field before it can be declared a solution for anything.
Yep, I think that's a very peculiar aspect of Matrix, about how it's run and why I have such a hard time trusting and recommending it. It's uncommon for opensource projects (especially "essential" ones) that adoption and fame must precede stabilization, as a condition to get to keep the cashflow and the lights on.
I don't think Matrix, starting up on venture capital, with an original but completely unproven idea, downplaying alternatives with FUD and superlative marketing, over-promising and constantly deflecting, was good community building.
Had they kept a lower profile and not an antagonizing one, they could probably have built and integrated better with the other communities in this space (and I'm not just talking about XMPP, which was on the receiving end of the FUD, but also about libera.chat whose OPs are right to be fed-up with NV).
Arathorn, 2023-09-23:
I don't know any admin who considers federation problems to be solved, and I don't think Matrix 2.0 to be ready for the general public, so I call this denial, but heh, this is subjective :)