Markdown Version | Session Recording

Session Date/Time: 30 Mar 2023 00:30

mimi

Summary

This meeting of the More Instant Messaging Interoperability (mimi) working group focused on defining the scope and prioritizing the work items as defined in the charter. Discussions centered around content format, message transfer protocols, identifier naming conventions, and end-to-end identity, with an emphasis on the relationship between these components. The meeting included presentations on content format, matrix as a transport, a Bd Transfer protocol, a Bb delivery service, and requirements for transport protocols. A significant portion of the meeting was dedicated to discussing requirements emerging from the various drafts and what the protocol needs in general, including aspects like naming and discovery, and consent.

Key Discussion Points

• Content Format:
  • The need for a standard format for common instant messaging features inside end-to-end encryption was discussed.
  • Debate arose regarding the inclusion of sender and receiver identifiers within the message container versus relying on the underlying MLS layer.
  • Concerns were raised about message editing and deletion policies, particularly regarding legal requirements to preserve evidence.
  • The distinction between threads and replies was extensively debated, including user interface and semantic implications.
• Message Transfer Protocols (Transport):
  • Presentations covered Matrix, Bd Transfer, and Bb Delivery Service as potential transport protocols.
  • The need to specify a server-to-server protocol was emphasized, with debate on whether a client-to-server protocol should also be standardized.
  • The concept of an "owning provider" and its role in state synchronization was discussed.
  • The importance of considering state consistency and fault tolerance in message transfer was highlighted.
• Identifier Naming Conventions and Discovery:
  • The distinction between System Specific Identifiers (SSI) and System Independent Identifiers (SII) was introduced.
  • Various approaches to identifier discovery were considered, including supporting only SIIs, building discovery separately, or using out-of-band mechanisms.
  • The importance of balancing usability, security, and privacy in identifier discovery was stressed.
• Consent:
  • Different consent models for message delivery were explored, including allowing messages to be sent without prior consent (but with potential spam concerns) and requiring explicit consent before delivery.
  • The need for mechanisms to handle spam and abuse in federated messaging systems was acknowledged.
  • The complexity of designing a consent system that accommodates different provider policies and user preferences was highlighted.

Decisions and Action Items

• Content Format Draft Adoption: The working group indicated strong support for adopting the content format draft, with the understanding that feedback and issues would be addressed through GitHub issues and the mailing list.
• Interim Meetings: The chairs proposed a series of regular virtual interim meetings to address the many weighty topics with deadlines.
• List Discussion: The chairs want a series of separate threads created on the mailing list to discuss the unresolved topics that were brought up in the meeting, and for the authors to get together.

Next Steps

• Rowan May will resubmit the content format draft for adoption in the working group's git repo.
• Start separate threads on the mailing list to continue the discussion on consent and discovery.
• Authors to get together and discuss a forward path for unresolved technical topics.
• Chairs to schedule regular virtual interim meetings to facilitate progress.
• Encourage contributions for alternative models for federation.