roadmap

(see previous personal wishlist: %9W+HJ7f...)

measure twice, cut once

minimum viable core

what are the fundamentals of the Scuttlebutt protocol?

1. capture the existing Scuttlebutt status quo
   • what is the problem space we are exploring?
   • what are our orthogonal solutions to cover the problem space?
   • what design hypotheses are our solutions trying to test?
   • what existing prior art exists to solve similar problems?
   • what have we learned so far from our exploration?
   • do we want to change any fundamental design decisions?
   • do we need to adapt any design details?
2. design a complete map for the next core Scuttlebutt protocol
3. write a thorough specifications for each part of the next core Scuttlebutt protocol
4. document accessible details of the next #ssbrs core protocol implementation
5. write maintainable tests for the next #ssbrs core protocol implementation
6. implement the next #ssbrs core protocol implementation
   • note: async modules may be delayed while the story for async Rust (async / await, futures) is in limbo.

at this stage we MUST make hard decisions in order to limit our scope: what features are must have versus what features are nice to have.

we SHOULD focus on changes that address well-known real-world problems with our current stack, or converging on better solutions given our learnings from exploring the problem space; we SHOULD NOT focus on changes for the sake of open-ended experimentation, or diverging from our current solutions to explore new territories within a larger problem space.

as a straw proposal for the next core protocol stack:

• data codec (#hsdt)
• multiplexing system (#bpmux)
• handshake and streaming cypher (secret-handshake, box-stream)
• private messages (private-box)
• database (flume logs and views)
• replication gossip (epidemic broadcast tree)
• flooded gossip (simple blobs, out-of-order messages)
• gossip scheduler (%HJTzjmTcr8FsKt34wmvN5ett4imneEzHp7dCab2rGpg=.sha256)
• minimal core rpc (minsbot)

features:

• ssb:// url identifiers (urlsafe base64)
• off-chain message content #offchain-content / #blob-content
• friendly bindings to re-use Rust bindings in other languages
• gracefully upgrade core message formats, typed contents, etc
• gracefully upgrade crypto keys, hash functions, etc
• gracefully change data codecs, transports protocols, replication protocols, etc
• be able to delete a post or feed

resources:

• scuttlebutt.rs website
• how to contribute to #ssbrs (shared opinions, processes, etc)
• how to use the #ssbrs cli

packages:

• ssb cli client
• ssb cli server

this milestone SHOULD NOT become a moving target. in order for a proposal to be accepted into the minimum viable core milestone, it SHOULD be convergent. any proposals that fail to converge within a reasonable time will be re-targeted for the next milestone.

next generation core

as a straw proposal for the next generation protocol stack:

modules:

• rpc plugins
• user invites (%LMYARKcJ5/HVrkfyGo0oSV4j/whFmpeiJlQnvPw53PE=.sha256)
• p2p relay (%Ci9CLOEmqt1/7ghJkHV76jzdWbB1cQxa92os7AShUB4=.sha256)
• multiple devices (#sameas)
• (private) groups
• dat blobs

packages:

• ssb pub server
  • external viewer
  • admin interface (stats, invites, etc)
• ssb mobile app
• ssb desktop app

resources:

• how to run an #ssbrs pub server
• how to build a mobile app on #ssbrs
• how to build a desktop app on #ssbrs

features:

• moderation
  • be able to block a feed (or block via following a list of blocked feeds)
  • flags: %QRrL2wc...
• reactions: %+r7V7cU...
• i18n + l10n for user interfaces