ietf-wg-mimi / mimi-protocol

Combined output from the MIMI Design Team
Other
4 stars 2 forks source link

Figure out versioning #52

Open rohan-wire opened 7 months ago

rohan-wire commented 7 months ago

Figure out if versioning of endpoints includes different options inside the same framing, or a whole new endpoint, discovered by the directory, or something else.

tgeoghegan commented 4 months ago

WG review: We don't yet really understand what the requirements are here. Do we want to allow in-band protocol version disparities?

Potential requirement: do all the providers participating in a room need to support the same version?

turt2live commented 4 months ago

I haven't yet caught up on the discussion at the interim, so apologies if I've missed something here.

The protocol may have multiple disparate versions associated with it:

At the transport level there should be some support for versioning which is not bound to the room/group itself. Servers should be able to negotiate with each other the versions of endpoints without affecting the overall room operation (in most cases). Where a room behaviour requires a certain operation to take place, that would obviously have impact and need to be taken into consideration. For example, a non-MLS room may require certain transport considerations to exist, though within those considerations versioning can be applied to upgrade/change body encoding and similar.

Policy and user/client synchronization may not require explicit versioning in the protocol, though a concept of "this room is replaced by X" may need to exist. For example, formalizing what an MLS ReInit looks like in user experience terms, and reflecting that back as protocol capability.