Open Infrastructure

Matrix API documentations can be found at https://matrix.org/docs/api/client-server/

Note: Deleting a room is probably not supported using Matrix API.

Waiting for login facility to get fixed in our Matrix instance.

• You can make use of this: https://github.com/matrix-org/matrix-js-sdk/ (it backs matrix-react-sdk/riot-web).
• Right now Matrix has a lot of incomplete clients: https://matrix.org/clients/
  - It's kind of an issue: lot of yet-to-be amazing clients, which do not support everything (i.e. E2EE, communities). People try a build new things instead of improving existing ones. (FYI)

A note on design: as we are often saying "mattermost looks nicer" I would suggest that we create a client that uses the UI of mattermost, but the sdk of matrix. This way we would combine the best of both worlds and in the best case don't need to re-invent the wheel.

One tiny thing I'd add to the mattermost design is to use the displayname as the username, but append in a slightly smaller font the real username: Nico Schottelius (@nico:ungleich.ch) (part in brackets smaller, maybe more muted).

Otherwise we can stay for 99.99% on the mattermost design

• I think it wouldn't be the case that we can easily use mattermost's things especially JavaScript components (https://github.com/mattermost/mattermost-webapp/tree/master/components) in our Riot web app.
• Although, we may be able to use the CSS (https://github.com/mattermost/mattermost-webapp/tree/master/sass)
• We can also just use bootstrap or some similar UI framework.
• We need to use some frontend framework for easier manipulation of DOM. See VueJS (It is easier one and adoptions is increasing day by day).
• Don't know yet whether we can use the matrix-js-sdk (https://github.com/matrix-org/matrix-js-sdk/) for all use cases e.g I don't see any login/logout/signup flows as it was mentioned in the http API document (https://matrix.org/docs/api/client-server/#/). Anyway, we can use both.
• Mattermost is very polished and to come closer to it in functionality and stability would be quite a feat.
• I think it would take serious amount of time to make it more usable than current Riot app which might be not a very good idea in current situation.

• Not Completed Yet. I want to take a look at Riot Web to check whether we can fork it and improve it (by replacing the UI)) *

Can you complete the estimate these days?

Nico, I am confused about the estimate because it is frontend work and I didn't have much experience on frontend. But, it wouldn't be small effort. I think it would take > 1 month to even have a usable frontend client not taking into consideration the pain points observed after actually running it in production.

Aye Aye Captain :)