RuView: Weekly Recap - Calibration Trust, Sensing Accuracy and Guarded Expansion
Three calibration and sensing fixes merged to bind receipts to room identity and correct telemetry, while bounded Realtek ingest landed as diagnostic-only and Neo4j integration stayed in draft. Pre-existing CI blockers around ADR numbering, rustls and firmware size were cleared.
Duration: PT3M10S
Episode overview
This episode is a short developer briefing from RuView.
It explains recent repository work in plain language.
- Show: RuView
- Published: 2026-09-21T09:44:04Z
- Audio duration: PT3M10S
Transcript excerpt
This excerpt keeps the crawler page concise. Listen to the episode or use the RSS feed for the full update.
Good morning from RuView, recapping development September 14th through the 21st, 2026.
9 pull request activity items, 3 additional commits this week.
The lead pattern: the project tightened trust around where sensing data comes from, before expanding what it connects to.
First, calibration is now bound to room identity. Merged work in PR 1903 adds process, session, room and model identifiers to the field calibration API, tracks the requested node set separately from observed contributors, and refuses to complete until all expected nodes report. Finish, promote, cancel and reset now…
Second, the sensing data path got more honest. PR 1944 brings bounded Realtek 8721 D X RAC 1 ingest onto current main, including the hardware parser, simulator, bounded server snapshot and status route. Fresh ESP32 data remains the primary source — unsigned Realtek data is explicitly diagnostic only and cannot…
Third, integration paused while infrastructure was repaired. PR 1932 proposes Neo4j sensing persistence with an opt-in, room-scoped M C P read surface for Hermes, but remains draft with release blockers after security review. In contrast, PR 1945 merged fixes for three pre-existing C I blockers:…
Nearby episodes from RuView
- Documentation Update
- Calibration Gets Tested on Real Hardware
- Clearing the Path for RTL8721Dx Sensor Data
- Sensing Data You Can Trust
- Weekly Recap - Making Real Hardware Tell the Truth
- Presence Detection Gets Honest
- Real Hardware Breaks the Simulator's Assumptions
- Firmware Onboarding Still Isn't Done