Aether HF 0.2.0-beta.159

0.2.0-beta.159 — 2026-11-14

New

• model: the FM channel model, a radio at each end (ADR-0115, step 10.2a)
• tools: the FM radio bench for step 10.2b, checked on the model

Fixed

• panel: the header keeps to one row; each view keeps its window size
• panel: a tab opened from Compact or Mini goes to Full

Documentation

• adr: ADR-0115, the FM channel model (FM-3), proposed
• adr: ADR-0116 and ADR-0117, the FM waveform and calibration (FM-4, FM-5), proposed
• adr: ADR-0118 to ADR-0125, Aether FM's remaining decisions (FM-6 to FM-13), proposed
• claude: Aether FM's model, bench and ADRs, in the working notes
• research: the FM bench's setup, for a TM-D710 and an IC-9700
• research: the FM bench's all-N kit, with its levels
• research: the HP 8496A as the FM bench's step attenuator
• research: check the FM bench's fixed pad against the step attenuator
• research: the FM bench with fixed N pads only
• research: two 10 dB pads in place of the 20 dB one
• adr: the author confirms FM's regulatory readings

Build and release

• nightly: fail, and email, only for what is serious

Internal

• release: 0.2.0-beta.159

Aether HF 0.2.0-beta.158

0.2.0-beta.158 — 2026-11-14

New

• control: a program may say hello, and the header names it in green

Documentation

• research: the Aether FM proposal, and Phase 10 started
• adr: Aether FM's first two ADRs, the product and the PHY boundary
• adr: ADR-0113, the PHY boundary, accepted for building
• adr: ADR-0113 built, step 10.1's gate passed

Internal

• aetherd: gather HF's physical layer into HfAir
• aetherd: the air takes each captured block and owns the busy detector
• aetherd: the station holds one Air, and an identity check proves it unchanged
• release: 0.2.0-beta.158

Nightly 0.2.1-nightly.20261114.abe8bf9

Unreleased

New

• model: the FM channel model, a radio at each end (ADR-0115, step 10.2a)
• tools: the FM radio bench for step 10.2b, checked on the model

Documentation

• adr: ADR-0115, the FM channel model (FM-3), proposed
• adr: ADR-0116 and ADR-0117, the FM waveform and calibration (FM-4, FM-5), proposed
• adr: ADR-0118 to ADR-0125, Aether FM's remaining decisions (FM-6 to FM-13), proposed
• claude: Aether FM's model, bench and ADRs, in the working notes

Build and release

• nightly: fail, and email, only for what is serious

Aether HF 0.2.0-beta.157

0.2.0-beta.157 — 2026-11-13

New

• link: the third rate review's try candidate, as bench-only switches
• link: the third rate review's G1, the prior's margin, as a bench switch
• link: the third rate review's G2b, the family offset's gain, as a bench switch
• link: the rate review's values on the 2300 Hz air alone, a bench switch
• link: the 2300 Hz air guesses closer and learns faster (ADR-0111)

Documentation

• research: the strong-path oracle on today's C1
• research: today's fixed rungs on the three strong paths
• research: the third rate review's Gate 2 draft, candidates and criteria
• research: the third rate review's Gate 1 survey
• research: Gate 2's draft cites the survey
• research: the author's answers to the third rate review's Gate 2
• research: Gate 3, the try candidate fails
• research: Gate 3, G1 at 1 dB passes the speed target and fails NVIS
• research: Gate 3, G1 at 1 dB with G2b still fails NVIS
• research: Gate 3, the 2300 Hz-only margin passes K1, K3 and K8
• research: Gate 3, the fresh draws of the 2300 Hz-only margin
• research: Gate 3, the last checks on the 2300 Hz-only margin

Build and release

• oracle: pin only the rungs asked for, per air
• the rate review's last checks on a candidate, brought back

Internal

• release: 0.2.0-beta.157

Aether HF 0.2.0-beta.156

0.2.0-beta.156 — 2026-11-13

Fixed

• aetherd: a connection let go from another thread ends on Windows too

Documentation

• research: the third rate review's Gate 0, the guess's evidence

Internal

• release: 0.2.0-beta.156

Aether HF 0.2.0-beta.155

0.2.0-beta.155 — 2026-11-13

Fixed

• daemon: a decompressed stream grows at most 100 times (ADR-0110)

Documentation

• research: the author's answers for the third rate review's Gate 0
• readme: on the air, with figures from the 80 m RadioPost sessions
• readme: what Aether drew on, and VARA as a black box

Internal

• harness: a Pat stand-in, and scenarios held to the wall clock
• fuzz: the link engine in a session, and the PHY receiver
• release: 0.2.0-beta.155

Aether HF 0.2.0-beta.154

0.2.0-beta.154 — 2026-11-13

Fixed

• host: BUFFER said again while the count stands still (ADR-0109)
• tools: rung_reading reads protocol 12's control frames, two_sides an open window
• aetherd: no read timeouts on servers' sockets, so Windows cannot end a quiet connection
• tools: rung_reading keeps blocks apart past 256 data frames
• aetherd: the keyed-audio watch listens in 1–2.5 kHz, under no interface's hum

Documentation

• field: the mobile's RadioPost mail to the gateway on 80 m, 2026-11-12
• research: the rate controller's third review, how C1 guesses a rung it has not measured

Internal

• fuzz: cargo-fuzz targets for the parsers that take outside bytes
• release: 0.2.0-beta.152
• release: 0.2.0-beta.153
• release: 0.2.0-beta.154

Aether HF 0.2.0-beta.151

0.2.0-beta.151 — 2026-11-12

New

• aetherd: debug mode sends every program's sessions, RadioPost's too

Fixed

• panel: the programs chip says "listening", not "not answering"

Internal

• aetherd: a session's notes keep its owner's kind, not two flags
• release: 0.2.0-beta.151

Aether HF 0.2.0-beta.150

0.2.0-beta.150 — 2026-11-12

New

• aetherd: say why a control connection ended
• aetherd: every program speaks for the station, and Setup says whose callsign it is (ADR-0108)
• aetherd: a program's own log goes in the shared zips (share.include)

Fixed

• panel: say plainly what a disconnect waits for

Internal

• host: wait for the whole payload, not its first piece
• release: 0.2.0-beta.148
• release: 0.2.0-beta.149
• release: 0.2.0-beta.150

Aether HF 0.2.0-beta.147

0.2.0-beta.147 — 2026-11-12

New

• aetherd: under automatic control only a program answers calls, and a program's session goes with it (ADR-0107)

Fixed

• panel: say how to set the rules under Host keying, and that RadioPost skips the host port
• config: refuse two of the station's ports on one number

Documentation

• readme: RadioPost, several programs on one station, and the receive filter
• manual: the host programs port is never 8515, and RadioPost needs Aether keying

Internal

• release: 0.2.0-beta.147