| 12 Aug 2025 |
 K900 | I don't think I am, honestly | 14:42:44 |
| K900 | My concern with Matrix as an ecosystem has always been, uh, for the lack of a better word, Matthew | 14:43:04 |
| K900 | What I'm trying to say is that it seems there's no real interest in fixing the core of Matrix from the side of Element | 14:44:10 |
| K900 | Who, for better or worse, de facto own the core of Matrix | 14:44:18 |
| K900 | And if you're going to do a full protocol redesign, doing it within current governance feels like it'll just be harder for no real reason | 14:45:13 |
 Charles | MSCs require a certain number of approvals from SCT members and the amount of SCT members that are element employees are such that element can fully steamroll any particular MSC or stall any particular MSC | 14:45:24 |
 QuadRadical (Ping) | i agree matrix is bad, but idk if theres a better option rn | 14:45:48 |
| QuadRadical (Ping) | personally i think ill stick to matrix until theres a nicer federated protocol i can host | 14:46:18 |
| K900 | Depends on the tradeoffs you're willing to make | 14:46:24 |
 Gnuxie 💜🐝 | then why do the work that they did do recently and why also make changes to creators and the room id... | 14:47:47 |
| Charles | that's not a protocol change really | 14:47:59 |
| K900 | I don't think that's a fix | 14:48:07 |
| Charles | a protocol change would be to stop using delta-state CRDTs | 14:48:07 |
| K900 | It's a patch | 14:48:09 |
| K900 | But it doesn't solve the underlying issue | 14:48:15 |
| K900 | And arguably adds a bunch of new fun issues | 14:48:24 |
| Charles | i am highly doubtful they will do this | 14:48:28 |
| Gnuxie 💜🐝 | regardless of what you think it is or its effect it is a change that should be welcome considering that there haven't been many like that | 14:48:37 |
| K900 | I don't know what a real fix will involve, but it will probably involve some sort of consensus thing | 14:48:40 |
| Charles | so matrix is screwed forever | 14:48:43 |
| Charles | yeah one would hope | 14:49:06 |
| K900 | Because I don't think zero trusting this is actually possible | 14:49:17 |
| K900 | Like, the current design will always have the rollback problem because that's how CRDTs are | 14:49:48 |
| K900 | So you either need an entirely different way of doing state resolution or some sort of common trusted consensus source | 14:50:22 |
| Charles | yeah, basically your options for delta-state CRDTs are for operations to be permanent XOR convergent | 14:50:43 |
| Gnuxie 💜🐝 | well there will be state res V3 in 6months to 1 year. | 14:50:47 |
| K900 | Something I was kinda thinking about was doing a roughtime kind of thing to agree on event ordering that's more than toposort | 14:50:52 |
| K900 | And then you just can't fork state ever | 14:51:16 |
 emily | is this publicly announced anywhere? | 14:51:57 |
| Gnuxie 💜🐝 | of course it isn't x3 | 14:52:05 |
