5 ms·

What protocol? If you look at spec on their website[0] you'll see 4/5 of the protocols are still marked unstable. Take room state resolution as an example. It h

What protocol? If you look at spec on their website[0] you'll see 4/5 of the protocols are still marked unstable. Take room state resolution as an example. It has pretty important security implications and it's nontrivial in a distributed setting. The algorithm is not stabilized, and it comes with a warning in the specification: Warning: This section documents the state resolution algorithm as implemented by Synapse as of December 2017 (and therefore the de-facto Matrix protocol). However, this algorithm is known to have some problems. I really like the idea of Matrix, but it seems a little immature for a government to use as a standardized messaging protocol. [0] https://matrix.org/docs/spec/ https://matrix.org/docs/spec/ [1] https://matrix.org/docs/spec/server_server/unstable.html#room-state-resolution https://matrix.org/docs/spec/server_server/unstable.html#roo...

alexkavon 8y ago
Yep. Their docs inspire little hope that the protocol will be little more than niche. How to build on top of this protocol without requiring someone with domain specific knowledge is beyond what's offered. Cue XKCD protocols comic.
- Arathorn 8y ago
  Right now there are 4.5M registered users on the matrix.org HS alone, so I’m not sure niche is right ;)
  - tacoman 8y ago
    An interesting number would be how many GCM/APN devices are registered in the matrix.org sygnal server. All of the Matrix/Riot users I know use private, non-federated home servers, but still rely on the project's Sygnal server(s) for push notifications.
    - Qwertie 8y ago
      Phone number/email verification is also still centralized as no one has worked out any other way.
      - kqr 8y ago
        Only to the extent that identity servers themselves are centralised, and Matrix lets you run your own.
      - acct1771 8y ago
        Important to note you're not required to use any of these features....considering sign up takes 7 seconds within the app (and for security... Email is a weak point), I bypassed it.
        acct1771 8y ago
        Both are a weak point, when it comes down to it.
    - Arathorn 8y ago
      Sygnal doesn't track this, actually. The best figure i have is that right now the matrix.org HS has 125,209 registered pushers - but as sygnal doesn't store registrations in its DB, I don't have an easy way to say how many Riots on other homesevers are routed through it.
- Karrot_Kream 8y ago
  Matrix is gaining adoption throughout the open source world super quickly. There are screenshots, guides, and links to join all over the place. Given that, I'm really impressed that anyone looks at Matrix and thinks of it as "little more than niche".
  - geezerjay 8y ago
    > There are screenshots, guides, and links to join all over the place. Given that matrix describes itself as an open standard, neither of those aspects says much about the project other that there's either some goodwill or a PR department.
    - Arathorn 8y ago
      fwiw there's no PR or marketing department, unless you count me trying to field questions.
  - alexkavon 8y ago
    Just clarify since I've upset some members of its community. I'm speaking from a development perspective. Sure Riot might have a large amount of users online, but what portion of that is people developing on top of the protocol. As I originally stated the docs are less than stellar, considering there are many missing pieces. So other than the original dedicated developers building a few things for a large community, it would appear that on the development side that it is very difficult to get started. As in: I have to seek out someone with specific domain knowledge and ask hundreds of questions and take lots of notes in order to extend and use the matrix protocol properly. Which makes something designed to be open, little more than niche.
    - Arathorn 8y ago
      Totally agreed that the spec is not where it should be; it has suffered from lack of dev bandwidth after funding mess last year, as we had to shift folks from spec to Synapse work in order to keep up with the growth of the network. we have someone joining end of May who should have specific bandwidth to work on it as part of their day job; we have a goal to close all the gaps by Aug.
Arathorn 8y ago
We’re working towards a 1.0 atm, which the French govt and many others are counting on. In terms of state res not being well specified: https://matrix.org/_matrix/media/v1/download/jki.re/hSmkLkFGVUnrXjtYcjsCUsyS https://matrix.org/_matrix/media/v1/download/jki.re/hSmkLkFG... is pretty comprehensive and should be merged into the spec relatively soon. I say kudos to the French govt for backing a FOSS project which is still evolving :)
- bgorman 8y ago
  I don't mean to be disrespectful. The first time I heard of Matrix/Riot was at OSCOn 2015. What has happened since then,and why has it taken so long to stabilize?
  - Arathorn 8y ago
    We spent massive effort working on: * implementing E2E encryption and getting it audited, which turns out to be very hard in a decentralised environment * refining Riot as a flagship client and trying to get it to a Slack/Discord level of UX (still a work in progress) * scaling, as the 1st gen implementation (synapse) failed to keep up with the growth of the network * adding features like Jitsi video conferencing, Widgets, Communities, Home pages, Rich Text Editing, Read Receipts, Synchronised Read Markers and much much more. * dealing with a major funding crisis last year. In retrospect we probably should have spent more time finalising the current features rather than adding on new ones, and since the end of last year we've been feature-frozen in order to finish the stuff in hand and get to a 1.0. Fixing the issues which remain in federated state resolution is probably the hardest problem to solve before a 1.0, and I'm happy to say that over the last few months we've made major progress in doing so. Agreed that in an ideal world this would have all happened sooner, but turns out that decentralised open protocols are harder than centralised startups like Slack and friends. Who knew?
    - Arathorn 8y ago
      (oh, and we rewrote (almost) the whole thing in Go, too)

alexkavon 8y ago

Yep. Their docs inspire little hope that the protocol will be little more than niche. How to build on top of this protocol without requiring someone with domain specific knowledge is beyond what's offered. Cue XKCD protocols comic.

Arathorn 8y ago
Right now there are 4.5M registered users on the matrix.org HS alone, so I’m not sure niche is right ;)
- tacoman 8y ago
  An interesting number would be how many GCM/APN devices are registered in the matrix.org sygnal server. All of the Matrix/Riot users I know use private, non-federated home servers, but still rely on the project's Sygnal server(s) for push notifications.
  - Qwertie 8y ago
    Phone number/email verification is also still centralized as no one has worked out any other way.
    - kqr 8y ago
      Only to the extent that identity servers themselves are centralised, and Matrix lets you run your own.
    - acct1771 8y ago
      Important to note you're not required to use any of these features....considering sign up takes 7 seconds within the app (and for security... Email is a weak point), I bypassed it.
      - acct1771 8y ago
        Both are a weak point, when it comes down to it.
  - Arathorn 8y ago
    Sygnal doesn't track this, actually. The best figure i have is that right now the matrix.org HS has 125,209 registered pushers - but as sygnal doesn't store registrations in its DB, I don't have an easy way to say how many Riots on other homesevers are routed through it.
Karrot_Kream 8y ago
Matrix is gaining adoption throughout the open source world super quickly. There are screenshots, guides, and links to join all over the place. Given that, I'm really impressed that anyone looks at Matrix and thinks of it as "little more than niche".
- geezerjay 8y ago
  > There are screenshots, guides, and links to join all over the place. Given that matrix describes itself as an open standard, neither of those aspects says much about the project other that there's either some goodwill or a PR department.
  - Arathorn 8y ago
    fwiw there's no PR or marketing department, unless you count me trying to field questions.
- alexkavon 8y ago
  Just clarify since I've upset some members of its community. I'm speaking from a development perspective. Sure Riot might have a large amount of users online, but what portion of that is people developing on top of the protocol. As I originally stated the docs are less than stellar, considering there are many missing pieces. So other than the original dedicated developers building a few things for a large community, it would appear that on the development side that it is very difficult to get started. As in: I have to seek out someone with specific domain knowledge and ask hundreds of questions and take lots of notes in order to extend and use the matrix protocol properly. Which makes something designed to be open, little more than niche.
  - Arathorn 8y ago
    Totally agreed that the spec is not where it should be; it has suffered from lack of dev bandwidth after funding mess last year, as we had to shift folks from spec to Synapse work in order to keep up with the growth of the network. we have someone joining end of May who should have specific bandwidth to work on it as part of their day job; we have a goal to close all the gaps by Aug.

Arathorn 8y ago

We’re working towards a 1.0 atm, which the French govt and many others are counting on. In terms of state res not being well specified: https://matrix.org/_matrix/media/v1/download/jki.re/hSmkLkFGVUnrXjtYcjsCUsyS https://matrix.org/_matrix/media/v1/download/jki.re/hSmkLkFG... is pretty comprehensive and should be merged into the spec relatively soon. I say kudos to the French govt for backing a FOSS project which is still evolving :)

bgorman 8y ago
I don't mean to be disrespectful. The first time I heard of Matrix/Riot was at OSCOn 2015. What has happened since then,and why has it taken so long to stabilize?
- Arathorn 8y ago
  We spent massive effort working on: * implementing E2E encryption and getting it audited, which turns out to be very hard in a decentralised environment * refining Riot as a flagship client and trying to get it to a Slack/Discord level of UX (still a work in progress) * scaling, as the 1st gen implementation (synapse) failed to keep up with the growth of the network * adding features like Jitsi video conferencing, Widgets, Communities, Home pages, Rich Text Editing, Read Receipts, Synchronised Read Markers and much much more. * dealing with a major funding crisis last year. In retrospect we probably should have spent more time finalising the current features rather than adding on new ones, and since the end of last year we've been feature-frozen in order to finish the stuff in hand and get to a 1.0. Fixing the issues which remain in federated state resolution is probably the hardest problem to solve before a 1.0, and I'm happy to say that over the last few months we've made major progress in doing so. Agreed that in an ideal world this would have all happened sooner, but turns out that decentralised open protocols are harder than centralised startups like Slack and friends. Who knew?
  - Arathorn 8y ago
    (oh, and we rewrote (almost) the whole thing in Go, too)