2025_05_11.md

TinySSB | 2025-05-11

Agenda

• updates
• SSB poll spec
• spec work
  • test vectors
• T5

T5: https://lilygo.cc/products/t5-4-7-inch-e-paper-v2-3

Updates

Mix:

• spec work: had a call with Nano, agreed on an arc
  • questions:
    • LogId vs FeedId (naming question)
    • type 1 message signature << still TODO /cft asks back: why should there be a difference, wrt type 0?/
      • CFT: you only sign over the main-chain packet, verification is all on the main-chain packet, no side-chain packets needed
      • Nano: I thought you said you needed the side-chain for something - next DMX/ sig?
      • CFT: no we designed it so you don't
      • CFT: to construct the signature you need to have already chunked up (this defines what's in the main chain bit)
• rust implementation: started with Piet
  • repo: https://github.com/tinySSB/tiny-ssb-rs
  • questions
    • test vectors
    • where is bipf getting used
• had a play with a SenseCAP T1000e
  • started organising buying locally: https://mixmix.nz/p/poneke-mesh/

Kevin

• still interested to know how I can help

Nanomonkey

• Common LISP implementation
  • questions
    • timing of messages
  • repo: not yet
• playing around with T5
  • MIX: does it have a case?
    • Nano: yes, I don't know how to get it off
  • power
    • battery: 1500mAh
    • 88mA while running => 9-10 hours active use
    • powering - wireless or USB

CFT

• T5
  • working on input
    • UI update is currently quite slow (0.5-2s)
  • similar to T-watch
  • this device is a full peer (can author it's own messages)
  • writing the apps here in C++ (not JS)
  • it can still behave as a relay still
  • repo at https://github.com/tinySSB/esp32-firmware/tree/cft-T5s3pro

SSB Poll Spec

CFT: I saw the message, had some questions

• inspired by Loomio, what is important? (pollSSB spec says "data model")
  • Mix: probably more about the vote styles

Spec

work to do

• re-order the spec << mix + nano
• finalise terminology
• test vectors
• proof reading / user-testing
• language/env set: [x] C++ [x] Py [ ] Kotlin [x] Rust [?] JS

Nano:

• propose we follow the spec arc
• GoSET as first case to study? Then main log entries, then sidechain pkt (chunks), then "apps"

Vectors

• name what we're testing << nano + mix
• how we're encoding them << coordination
  • CFT: prefer to have a hex encoding (s-expressions from Rivest: #11223344# for a byte array)
    • https://en.wikipedia.org/wiki/S-expression
    • https://www.ietf.org/archive/id/draft-rivest-sexp-01.html (2023)
  • Nano: I think it should be inline in the docs
  • Mix: parts I think we want
    • description
    • input data (feed_id_1, feed_id_2, ...)
    • expected output data (xor, highest_feed_id)
• make them
• have everyone test them

ACTIONS / TODO

CFT

• produce some GOSET test-vectors (probably: from 0 to 5 FeedIDs), put them in a branch of the spec repo

Nano + Mix

• working on the spec!
• getting test-vectors requests to Christian

Mix

• call with Kevin running through the protocol

Kevin

• Proof reading the spec