Matrix Community Year in Review 2022
blog.neko.dev
blog.neko.dev
And in general I‘d love to get rid of sending read recipes but I remember there being a whole discussion around that.
Disclaimer: I'm one of the maintainer of NeoChat
The funding trouble is somewhat surprising. Has anyone gone through and done a deep dive comparison of other open source projects that are doing fine.
Mastodon, doing fine on donations because the product small and well defined ?
IPFS, from the start they planned to make money with fileCoin and IPFS is a supporting service they develop because it is useful for fileCoin?
Elasticsearch, well defined product with traditional enterprise feature split and users who are very accustomed to paying for this kind of a product ?
MongoDB, well defined product and much more restrictive license ?
Supabase, much less to develop because they are mostly gluing existing code together. But they are bringing out some great new things and the glue has value . And again its a database so people are accustomed to paying ?
gitlab, nginx...
There are actually countless open source companies doing well when they are in a category of product that people are accustomed to paying for.
How far away would matrix be from becoming something like Cloud Native Computing Foundation or Linux Foundation. I suppose larger tech companies need to start making money with Matrix not Goverments or smaller anti capitalism groups.
If my memory serves, Elasticsearch suffered a lot from Amazon offering their product for free without contributing in any manner. That's not very far from what is happening to Matrix.
I noticed that while reading it as well. The call for funding ap few weeks ago did not sound nearly as dire. I wonder how things are going with that. I had wanted to donate to Matrix a number of months ago, but there was no easy, non-Patreon way to do that. I guess I assumed that with the new ask for donations and setting up DonorBox would have a pretty large response.
It's also strange how quickly this situation developed. I read every TWIM (at least I did before RSS broke recently) and follow the project and some developers on Mastodon, and don't recall hearing anything about financial woes. If anything, it seemed like at least New Vector was doing well financially. It isn't clear to me whether the financial issues are with Matrix directly, or New Vector. I don't think this is mentioned for the two developers who were laid off either.
Overall, the tone and urgency of the 2022 update struck me very off guard - not what I was expecting.
This is a major shift from the past; perversely it turns out that the reward for the core team for having improved synapse’s scalability and performance and generally improved Matrix is to undermine our primary route of funding.
The initial “funding Matrix via the Matrix.org Foundation” blogpost got very little feedback (other than complaints from the community that it was biased against indie contributors participating in the foundation, which was not remotely the intention). Meanwhile it’s become increasingly apparent that while it’s viable for a 3rd party matrix vendor to build a sustainable business on top of Matrix… it’s not so viable if you are also shouldering the costs of almost all the underlying core Matrix dev. Hence trying to explain to folks unambiguously that we need a sustainable way to support the Foundation, and sending up distress flares in the xmas blog post.
I am very bemused that some of the reaction has been resentment about the blog post being “scary” and “unsettling”. I’m afraid that we default to transparency, and while this may seem abrupt, this is because we’ve shifted our assumption that large govt-scale deployments will support the core team by default (unless we explain very unambiguously how important it is.) Sorry if this upset anyone’s christmas, but imagine how the core team feels (not to mention those whose jobs were impacted).
Weirdly, there's always a ton of netagativity and pessimism around Matrix on here, so I did just want to say I am a huge fan of the work, and general approach of the organisation, and as an avid user, am happy to contribute back financially. There's certainly no resentment on my part. If anything, I'm just concerned.
If they market towards companies - the search is still very bad, doesn't give any options (like search within a timeframe, search types of attachments etc... compare that to discord or slack e.g.), it's completly missing on mobile.. and generally sucks at showing attachments - seeing something from 3 months ago is incredibly annoying if you have hundreds of attachments, sometimes due to a bug even impossible. Hopefully slide sync and other stuff will adress all this filtering, searching and moving around a timeline quickly (going back a week ago? good luck). Also the amount of bugs as of late is quite worrying - after a year or two I am having issues with encryption all over the place, the notifications were buggy (maybe the version that's not showing up for me in the playstore already fixes that)... so I feel like the resources should be focused towards polishing and not rewriting everything again.
And if gamers and home users are their niche and want to target those users - discord users were asking for custom emotes for years now and while a solution for stickers exist, it's not for emotes.
Also I don't understand why pouring resources into its own videostreaming when jitsi seemed like a very good option already - again in my opinion should've been a low priority and scheduled for much later in the pipeline.
Again I truly hope Matrix succeeds, but making a better experience, useability and polished product should be the focus now. While third party clients may have other features, it's Element that will be used the most by people, at the start at least.
It’s the first rewrite of the iOS app, which was 8 years old. And it unifies the android and iOS codebases at last, to stop having to implement everything twice on mobile.
> (that has to be wasting a lot of developer time).
no, it is already spectacularly reducing dev time; basically halving it.
You can (re)read the explanation at https://matrix.org/blog/2022/12/25/the-matrix-holiday-update...
> And secondly they seem to be kinda stuck in terms of its vision.
The vision is pretty simple: radically fix perf, fix bugs by converging SDK, then add new features like encrypted search. Native VoIP is critical, given Jitsi’s encryption is beta and doesn’t interop with Matrix’s, nor is it decentralised, and we’d rather invest time otherwise spent maintaining large-scale Jitsi deployments making Matrix support VoIP out of the box.
In terms of the recent Android notification bugs: yes, they should be fixed in 1.5.16, released 2 days ago. And they are precisely the sort of bugs we are addressing by unifying the mobile apps onto a single well-tested SDK.
Not really sure about iOS, but I was thinking of android - there was the old element(riot), the current version(riotx?) and now the proof of concept.
> You can (re)read the explanation at https://matrix.org/blog/2022/12/25/the-matrix-holiday-update...
Sorry but I am still very confused about the naming, what is Element X, what is Element R. I am assuming Element web and mobile are the current stable versions but what's the difference between R and X is still not clear to me.
> The vision is pretty simple: radically fix perf
How? Are you going to keep developing both synapse and dendrite or are you planning to sunset synapse eventually and only develop dendrite? How close is dendrite to being stable and any plans for making migration guides between the two?
Element X is the new app, which implements the new faster sync and is built for performance from the outset.
Element R is an upgrade to the existing separate apps, which converges just the encryption on the new rust implementation, to minimise future encryption bugs by sharing the same code over all 3 platforms.
In terms of fixing perf, the solution is to switch to the new sync API (which will be supported on both synapse and dendrite) and use the new client codebase. Synapse and Dendrite are relatively similar in performance in comparison. Synapse dev will continue; Dendrite dev will continue focusing on embedded servers. Dendrite is very close to being stable, and yes there will be migration guides.
The relevant section:
> The interesting plot twist is that Dendrite development has ended up increasingly focusing on embedded matrix server use cases - particularly to power Peer-to-Peer Matrix, where clients require a server to be embedded within them. So while Synapse has ended up increasingly focusing on large-scale deployments, Dendrite has ended up pursuing smaller instances (which is ironic, given originally it was meant to be the other way round!).
Whoa... how is that gonna work?
Liberapay may not be as easy as Pateron, but especially with Patreon's UX getting worse over time, the major difference for me is now that you usually pay/renew each donation yearly.
I've donated for a couple years with liberapay and never had a problem doing non-recurring donations.
Any public roadmap for that?