Discussion: Social context working group

[wesbiggs]

We are convening a working group to collaborate on the DSNP architecture for contexts.

Contexts define discrete communities for social networking. There is an implicit "global context" in DSNP; all Broadcast Announcements (and their Replies and Reactions) form a global, public stream, where all DSNP Users have membership and agency, and there are no gatekeepers or system-wide moderators.

However, since the DSNP specification was first written, it has become increasingly clear that some of the goals of the protocol require supporting different definitions of communitites. Non-global contexts may have differing rules for membership and different local norms.

The first goal of the working group is to generate "version 1" requirements for contexts.

Here is an initial strawman proposal for v1 scope:

In scope for initial version:

• Context creation and state management
  • Naming and identifiers
    • Algorithm for globally unique identifier
  • Controller relationships
  • Context hierarchies
  • Moderation facilities
    • Moderator identification
    • Moderator capabilities
  • Membership definition
    • Allow lists / block lists
    • Rule-based access
• Application to DSNP content
  • Broadcast Announcements
  • Replies/Reactions

Out of scope for initial version:

• Private contexts (i.e. encrypted content)
  • While important, we will treat key management concerns as orthogonal.
• Context-specific user identity
  • We assume for now that participants are free to create new/additional DSNP User Ids if they desire context participation to not allow correlation.

[JoeCap08055]

My only comment is that, while private contexts/encrypted content is orthogonal to this discussion in detail, the discussion should be conducted with an awareness of some basic foreseen requirements for private content, so as not to design a protocol for social context that can't easily be enhanced to support private content.