A ships receiver going live is news too
When an ADS-B receiver delivers its first packets to Skydex, three things happen within a minute. Its owner gets a bell saying the line they pasted worked. The #network channel gets a card with the station's name, roughly where it listens and who runs it. And the live feed inside the game gets a row, because a receiver joining is the one piece of network news that reads as growth rather than plumbing.
None of that happened for a ships receiver. Not once.
How the gap got there
All three announcements were wired into the sweep that runs every minute over the aircraft ingest. That sweep's entire input is the list of clients connected to it, and an AIS station never appears in that list: its id rides as a Bearer token in an AIS-catcher flag, or as a tag block in a UDP datagram. It opens no Beast connection, so as far as the sweep is concerned it does not exist.
The water side did detect the moment. The code that credits a station for its delivery has always noticed the first one, and it wrote a line to the log. A log is not a surface anybody's owner reads.
Tonight made it concrete. One person registered two receivers a little over half an hour apart, an aircraft one and a ships one, both under the same name. The aircraft receiver produced two feed rows and a bell. The ships receiver produced nothing at all, while feeding perfectly well.
What changed
The announcement moved into a single place both bands call, so the two can no longer drift. An AIS station's first message now rings the bell, posts the card and writes the feed row, exactly as an aircraft station's does. The only difference is the footer on the card, which names the network the receiver actually belongs to.
Two smaller things were wrong in the same corner, and both are fixed.
"Your receiver is back" could never reach a ships station. The silence ladder works on both bands: go quiet for twelve hours and you get a bell, a week and you get mail, a month and the station moves to the dormant list. The other half of that pair asked the aircraft client list who was delivering, so a ships receiver could be told it had gone quiet and could never be told it had come back. Its outage flag would have stayed set for good. Delivery is now proven the way each band can prove it.
The mail about a silent receiver printed the wrong instructions. A ships feeder reading "check the dongle, readsb stays active with a frozen message counter" and then a Beast connector line was being sent to debug hardware that has nothing to do with the receiver that stopped. Those messages now carry the AIS-catcher flag with the station's own id, the raw NMEA door as the alternative, and a symptom that matches the software the person is actually running.
Nothing here is visible to anybody who does not run a receiver, which is rather the point: the people it is visible to had been feeding us into a room with the lights off.