Hacker News Board Hacker News, browsed like a classic message board
Live data from Hacker News

Matrix v1.15

matrix.org

Post reply on HN
Print view View original

121–130 of 131 posts

Re: Matrix v1.15

#121

Post by BrenBarn »

mindslight wrote: post #119
BrenBarn wrote: post #115

Earlier quoted context omitted.

That's true, but in another sense I think it's just that trying to "do it all" encryption-wise is significantly harder than some people realize. Having encryption that's really "safe" raises barriers to casual use that most casual users aren't really willing to accept. Like suppose I use Matrix only on my phone, so I just have the one device. Then I lose my phone and have to get a new one. How do I regain access to m…

I'd say that all of those usability problems are made worse by bolting on encryption after the fact. It makes it so there are now two layers of authentication/identity, rather than a single one. Whereas with built in encryption, you can always punt on solving the problems you list and make implementations that are less secure. For example, just store the keys on the server, accessed by the password. That's obviously…

Yeah, I don't disagree. Just saying that even if they had put encryption in at the outset, there are still tough issues to solve in terms of integrating that with user expectations.

Re: Matrix v1.15

#122

Post by lrvick »

tabbott wrote: post #116
lrvick wrote: post #66

Earlier quoted context omitted.

Zulip lacks end to end encryption and decentralization so they are not remotely comparable.

If you're doing consumer chat with friends and family, I'd recommend Signal over Matrix or Zulip. For a community: A lot of communities want anyone in the public to be able to join their spaces and read their channels. For that use case, E2EE makes the chat system slower and less usable, with limited security benefits over using web standard encryption. What E2EE may protect you from is a malicious server operator re…

Signal is a non starter. Never had an account and barring major architecture changes, I never will. Weakly secured centralized control of metadata with no option to self host your way out of it, and a requirement of a phone number which makes it a non starter for friends in countries that have good reason to be anonymous.

Zulip is at least self hostable but all private conversations in plain text on the server and lock-in to only being able to access chats on the one server are also both non starters. It is not a matter of trusting the admin, it is trusting that the Linux server was deployed with no exploitable flaws, and the admin never gets a court order. Too much trust for me.

Zulip is maybe a nicer UX alternative to IRC, but it is not comparable to Matrix.

Re: Matrix v1.15

#123

Post by freilanzer »

Arathorn wrote: post #85
freilanzer wrote: post #83

I like the idea of matrix and element, but the execution is just so lackluster. Even Signal added a lot of quality of life improvements, and WhatsApp is consistently getting better. I don't like the idea of using a messenger owned by Meta, but honestly, messengers absolutely need an excellent UX and UI, QOL features, etc. and Element is just far behind in this regard.

$50 says that this is yet another instance of the failure-mode explained in https://news.ycombinator.com/item?id=44392479 .

ElemntX does not support all features that Element does, so if I'm using a space to group rooms, then I can't use it.

Re: Matrix v1.15

#124

Post by freilanzer »

Arathorn wrote: post #90
freilanzer wrote: post #82

Earlier quoted context omitted.

> One of the big problems is that folks judge Matrix based on the legacy Element apps, which have now been succeeded by Element X Yeah, wake me up when X supports all the features the original app has. Until it's usable, there is no point.

which features are you actually missing? the only gap now is threads (in beta), spaces (in final design) and widgets (which were very very rarely used). the whole "i can't use X until it has precisely the same featureset as Y" feels a bit spurious, especially as you can run each app side by side and revert to the old one if you really need a given feature.

Spaces.

Re: Matrix v1.15

#125

Post by sneak »

BrenBarn wrote: post #53
cyphar wrote: post #43

Earlier quoted context omitted.

Unfortunately, they still use libolm[1] for e2ee which is deprecated[2] and has known security issues[3]. The maintainers appear to not be interested in switching to the newer Rust-based library. Matrix argue that the timing channel attacks are not possible over a network, but the history of timing channel attacks argues that this very few protocols are this fortunate (most people thought timing attacks against TLS w…

Yeah, although Matrix is theoretically about being an open protocol supporting a range of clients and servers, in practice it winds up being heavily skewed to just Element/synapse. I think this is partly because there is still too much churn in the protocol. A decent amount of that churn is improving things, but it still makes it too hard for average-joe devs to keep up with what's hip. I don't think there's much cha…

AFAIK none of that churn is around functionality that actually matters that much to end users, however. Certainly nothing as important as “working clients”.

Re: Matrix v1.15

#126

Post by WhyNotHugo »

godelski wrote: post #45
deberon wrote: post #39

There’s a surprising amount of negativity in here. Matrix is great for me. I’m happy to see the improvements. I’m also happy to hear the team is approaching sustainability. Matrix isn’t perfect but it has only improved in the 5 years I’ve been using it. I’m looking forward to what they can do in another 5 years.

I think HN just tends to be negative, though it seems to be getting more negative. I think a lot of people don't understand that finding problems is easy. Fixing them is hard. Which yes, you need to identify first but criticize, don't complain. But it's funny because anytime someone talks about Signal lots of people point to Matrix as "better". I'm not going to do the reverse because they solve different problems, wh…

Competition can be good, for example, if we had new clients for existing federated IM protocols.

But Matrix brings fragmentation more than anything else: a brand new protocol that's slowly re-inventing what we already had. Matrix clients don't really compete with existing ones for open (or closed) protocols, they exist in an entirely new sphere.

I don't think this is a net positive; the general ecosystem has finite manpower, the more we fragment it, the less each protocol gets. In the end, we have a new federated IM protocol which doesn't bring many new things to the table, but we still don't have any reliably easy-to-use clients.

Re: Matrix v1.15

#127

Post by tcfhgj »

WhyNotHugo wrote: post #126
godelski wrote: post #45

Earlier quoted context omitted.

I think HN just tends to be negative, though it seems to be getting more negative. I think a lot of people don't understand that finding problems is easy. Fixing them is hard. Which yes, you need to identify first but criticize, don't complain. But it's funny because anytime someone talks about Signal lots of people point to Matrix as "better". I'm not going to do the reverse because they solve different problems, wh…

Competition can be good, for example, if we had new clients for existing federated IM protocols. But Matrix brings fragmentation more than anything else: a brand new protocol that's slowly re-inventing what we already had. Matrix clients don't really compete with existing ones for open (or closed) protocols, they exist in an entirely new sphere. I don't think this is a net positive; the general ecosystem has finite m…

> But Matrix brings fragmentation more than anything else: a brand new protocol that's slowly re-inventing what we already had.

We certainly didn't have decentralized rooms before

Re: Matrix v1.15

#128

Post by v3ss0n »

lrvick wrote: post #66
v3ss0n wrote: post #50

Earlier quoted context omitted.

When compare to Zulip , Matrix usability is nothing close to that.

Zulip lacks end to end encryption and decentralization so they are not remotely comparable.

For internal company uses nobody needs that

Re: Matrix v1.15

#129

Post by v3ss0n »

lrvick wrote: post #122
tabbott wrote: post #116

Earlier quoted context omitted.

If you're doing consumer chat with friends and family, I'd recommend Signal over Matrix or Zulip. For a community: A lot of communities want anyone in the public to be able to join their spaces and read their channels. For that use case, E2EE makes the chat system slower and less usable, with limited security benefits over using web standard encryption. What E2EE may protect you from is a malicious server operator re…

Signal is a non starter. Never had an account and barring major architecture changes, I never will. Weakly secured centralized control of metadata with no option to self host your way out of it, and a requirement of a phone number which makes it a non starter for friends in countries that have good reason to be anonymous. Zulip is at least self hostable but all private conversations in plain text on the server and lo…

Zulip's topic focus chatting get the work done and kept information organized, very easy to search and assign tasks. That's huge upgrade over IRC since information is organized at the protocol level.

i agree matrix is desirable to replace slack only in case of where E2E is necessary.

For companies, inspection of chat logs may needed by top management in case of dramas.

Re: Matrix v1.15

#130

Post by makematrixriot »

So my issues are with Element and not matrix specifically.

Dropping Element One was not a great decision. This was a great opportunity to get funding and let hobbyists in a the ground level without insane business pricing that reminds you of Microsoft or Atlassian.

The devs convinced the leadership to completely halt improvement of features and implementation of new features to rewrite their codebase in an obscure and niche language that offers minimal benefits. Yes Rust is cool. Yes rust may be more secure. Completely re-implementing your main client in a new language while removing tons of functionality was a bad decision.

There's no mobile client support for spaces, you can't browse public rooms from mobile, threads is horrifically broken, replies are terrible, they completely redesigned the emojis but dropped support for / commands.

Element Server Suite is terrible, it isn't even close to enterprise level. There's no meaningful way to manage rooms at an admin level. There's no way to manager users through MAS in a reasonable way. You can't let admins take over rooms unless there is an admin bot in every room, including DMs. There's no way to manage federation. Requiring K8S to run a chat server is insane. The entire stack is way too heavy and most of the features that would make it worthwhile are pure vaporware.

Post reply on HN