Room Summaries (was: Calculate room names server-side)

[matrixbot]

Documentation: https://docs.google.com/document/d/11i14UI1cUz-OJ0knD5BFu7fmT6Fo327zvMYqfSAR7xs/edit#
Author: @ara4n
Date: 2018-06-10

The rules for calculating the display name of a room are complicated. It would be preferable to offload some of this work to the server.

This becomes even more important with the advent of lazy-loaded members (#1227), as a given client will typically not have enough info to correctly name the room otherwise.

This would probably mean adding a field to the /sync response.

Care would need to be taken to make sure that the textual parts of the name can be translated into the user's preferred language. This would probably mean providing the information in a structured manner (possibly as well as supplying preformatted strings in one or more languages for clients to use for simplicity if they prefer).

(Imported from https://matrix.org/jira/browse/SPEC-425)

(Reported by @richvdh)

[matrixbot]

Jira watchers: @richvdh

[matrixbot]

Links exported from Jira:

relates to https://github.com/matrix-org/matrix-doc/issues/573
relates to SYN-265

[ara4n]

This comes up as a blocker for lazyloading members (#1227), where otherwise we don't have enough data to calculate the name locally. Going to write a proposal for how to fix this (which I vaguely remember sketching out already somewhere, but i don't think has ever been actually written down).

[KitsuneRal]

Since room versioning MSC is apparently overtaking this one in the race to being specced - can we please include the room version in the summary so that clients wouldn't need to explicitly request m.room.create just to figure if they can even read the contents of that room?

[richvdh]

@KitsuneRal: let's take room versions back to #1425.

[richvdh]

This has been implemented in matrix-org/synapse#3574, so I'd suggest that anyone with objections makes them quickly.

[turt2live]

I vote that this MSC should enter the final comment period with the view of having it be a spec PR next week. Can the Spec Core Team either approve going into FCP or comment with any concerns. Those outside the team with concerns/suggestions are also welcome to raise them.

• @anoadragon453
• @ara4n
• @dbkr
• @erikjohnston
• @KitsuneRal
• @mujx
• @richvdh
• @turt2live
• @uhoreg

(also this has been in review for a while now, but no one updated the labels)

[ara4n]

Having been playing with it over the last few days on riot/web, it seems to be working well in practice. So 👍 from me.

[uhoreg]

Added a little question to the doc, and I'm a bit confused about the calculation of the room name (as with Manu's comment) but I'm not seeing anything fundamental that should stop it from going to FCP.

[richvdh]

In the same vein as #1227 (comment), I think we should declare that this had a de-facto FCP, and now needs speccing.

[turt2live]

Spec PR: #2035

[turt2live]

merged 🎉