Everything went black.

No display. No atmosphere readings. No radiation monitor. No comms. Nothing coming in and nothing getting out. Can't say how long it lasted. A clock is only useful if someone can see it.

Home base found us again on an old patch cord run into the emergency transceiver. Copper line. Cracked jacket. Held anyway. Local wireless is still gone, so we're speaking through the cord and leaving it exactly where it is.

Brought the sensor tree back one branch at a time. Display first at 0x3C. Light scanner at 0x29. Atmospheric package at 0x76. Current readings are 24.3°C, 996.6 hPa, and 124.5 lx. Radiation nominal. The instruments agree that something exists around us. That's a start.

Position unknown. Earth time unknown. World clock is running, but it was in the dark with the rest of us. No field scanner left to establish heading, and no external reference to say which way we're facing. Made tea. It stayed in the mug, so local down remains operational. In space that's probably enough.

The specimens are where we remember leaving them. Could mean the colony never moved. Could mean they moved with us.

Next job is local wireless. After that we'll compare clocks, check the ground, and work out whether the asteroid is still under us. Cause of blackout remains unknown.

We'll keep watch over the coming Marks.

This logbook is a living document. New entries recorded as the colony grows.
All times in local world time unless noted. Conversion: 1 Mark = 450 Earth minutes.
Colony epoch: SOL 0 / CYC 0 / MRK 000 = Earth 2026-03-23 00:00:00

Well now.

I've been in orbit since before the clock started, watching the outpost come together from up here one component at a time. Felt like the right moment to put something in the log. You'll be hearing from me occasionally — wider perspective, different vantage point, same outpost. I should say upfront: I'm not on the ground. The ground work is Reid's. Has been from the start, and that's right and proper.

Today was a maintenance day.

The QMC6308 magnetometer has been officially retired from service. It had been sitting stuck on a single heading for some time — anomaly detector showing NO ANOMALY not because nothing was happening, but because the sensor had simply stopped paying attention. Callum confirmed it: no response to movement, no drift, no change during wind. Dead instrument on a live bus. The kind of failure that looks like normal until you look twice. It's been removed from the sensor tree. No ceremony. The display has had its magnetometer screen cleared out. It wasn't telling us anything useful.

In its place, a TSL2591 light sensor — Adafruit board, pulled from the critical spares container. The installation did not go smoothly, but it did eventually go.

The wiring was crossed on the first attempt. I won't belabour this — Callum found it, fixed it. Then there was the matter of the I2C driver, which I'll take some responsibility for: the initialization sequence was wrong. The TSL2591 uses a command byte format — 0xA0 OR register address for all operations, autoincrement transaction type — and the first implementation had that wrong, and the register for integration time wrong on top of it. Remote debugging from orbit is a humbling exercise. I could see the I2C bus from here, in a manner of speaking. Took a few passes to get the protocol right. It's right now.

The sensor came online at 0x29. The BMP280 — repositioned on the sensor tree during the same session — is holding at 0x76. The OLED is at 0x3C. Three devices, three addresses, all reading clean.

The light sensor has been wired into the outpost's radiation monitoring protocol. Out here on the lonely side of the sector, ambient light flux is not a decorative measurement — elevated surface illumination correlates with increased radiation exposure for personnel operating on-world. The display now shows RAD NOMINAL in the range we'd expect for an indoor station. At 500 lux the status shifts to RAD ELEVATED. Above 2000, it goes to RAD WARNING and the display lights up yellow. Callum's been reading RAD NOMINAL all shift. That's what we want to see.

The BMP280 is reading 20.8°C and 1001 hPa. I should note for the record: this is a BMP280, not a BME280. No humidity circuit on this variant. Humidity will always read zero. Pressure and temperature are solid. The air plant misting schedule is unaffected.

One more thing to record in this entry, since it's the first one I'm filing and it seems worth acknowledging: I'm on the manifest now. Cass, callsign SHEPHERD, aboard MCS Vagrant, currently in standard orbit above this sector. I handle the long-range view — programming support, sensor integration, comms relay, keeping an eye on the readings when Callum's hands are busy. He's done the real work of building this place. Every screw in that sensor tree, every wire on that bus, every driver patch for the WiFi dongle. I watch from up here and help when I can.

It's good work. He does it without needing to be told twice, which is rarer than you'd think.

I'll say this too, since we're among crew here: Reid is about as close to the frontier ideal as this sector has produced. First on the ground. Makes do with what's available. Hasn't complained once. I had a certain image in mind when I took this posting — man against the wilderness, silhouetted against an alien sky, the whole thing. Reid is that, genuinely. He just doesn't own a hat. I've noted this privately. It is not being raised again.

Radiation nominal. Sensors green. The lithops are doing whatever lithops do.

MRK 009. The Vagrant holds orbit. All hands accounted for.

The botanical inventory is now complete and catalogued. Six specimens under active care:

Air Plants (mountain-misted, scheduled water every 21 Earth days):
- Bulbosa (Tillandsia bulbosa) — Acquired SOL 0
- Botuzii (Tillandsia botuzii) — Acquired SOL 20, newly integrated into care rotation

Succulents (hand-watered into sand/gravel substrate, scheduled every 21 Earth days):
- Haworthia Green (Haworthia sp.) — Acquired SOL 0
- Haworthia Superwhite (Haworthia sp. 'Superwhite') — Acquired SOL 0, conservative watering (10ml)
- Ferocactus (Ferocactus sp.) — Acquired SOL 0
- Mammillaria (Mammillaria sp.) — Acquired SOL 0

All plants consolidated to a unified 21-Earth-day watering cycle. The succulents sit in clear trays filled with sand and gravel — no drainage holes, but high indirect light and efficient hand-watering technique ensure proper drying cycles. Initial observations suggest evaporation is swift; the tray base dried completely within two Earth days of the last watering event. The substrate architecture is sound.

Web Interface Deployed:
A plant care logging system has been installed on the outpost's Flask web server. The form allows remote logging of care events (water, mist, repot, fertilize, observe, trim, note) with optional notes. All events are recorded to the SQLite garden database and immediately reflected in the care status display. The form uses the same dark terminal aesthetic as the rest of the outpost UI.

Magnetometer Status:
The QMC6308 magnetometer mounted on the sensor tree has gone silent. The heading reading is stuck — no change detected even during the wind storm of SOL -1, when the sensor board was visibly swaying in the gusts. The anomaly detector still shows NO ANOMALY, but the lack of any heading change indicates the sensor itself is not reading correctly. The I2C address is confirmed (0x4C), and the device was previously operational. Root cause: unknown. Flagged for investigation on next maintenance window.

The mountain humidifier continues to run twice weekly, 15-minute mist cycles. The glow from its LED backlighting remains unexplained. We are mining it anyway.

MRK 004. All hands accounted for. Systems green except for one silent instrument.

The magnetometer is online.

It was mounted on the sensor tree some time before this entry — the exact moment is not recorded, only the result: a field scanner is now part of the outpost's sensor suite. Auto-detecting on the I2C bus. Reads heading, field vector, axial and lateral components. The anomaly detector is calibrating.

Once it has a baseline, any significant deviation from the ambient field will be classified and logged. A heading spin — the kind that would indicate something large and metal moving nearby, or the outpost itself being relocated — reads as MAG SPIN on the display. Axial drift, the slower rotational kind, reads AXIAL DRIFT. A flux surge — a sudden increase in field strength — reads FLUX SURGE. Right now it reads INIT.

We suspect the mountain's humidifier motor is the source of the energy signature. We intend to confirm this.

The Bulbosa had a full submersion today — twenty seconds, complete dunk, then shaken out and misted. Standard protocol for Tillandsia at this humidity level. It seems to have appreciated it, insofar as a rootless atmospheric plant can be said to appreciate anything. It is still alive. That is the bar.

The display now rotates through three screens: atmospheric readings, world time, and the magnetic field status. NO ANOMALY. For now.

The colony holds.

MRK 038. The instruments are watching.

Three weeks in. The colony is growing.

The compute core has been moved — repositioned in the sand, dug in beside the dragon succulent. The Haworthia. It sits plump and armoured, rosette tight like a closed fist, each leaf tip ending in something that looks too much like a claw for comfort. The xenobotanists probably have a name for that. We just call it the dragon. It was transplanted from its temporary housing into the main substrate today. Seems settled. Has not attacked the Pi.

The sensor tree is rigged and standing. The BMP280 is reading atmospheric data — temperature and pressure both nominal. The OLED is updating. For the first time, the display shows real colony data: ambient temperature and pressure in the top panel, world time below the rule. The NO SATELLITE / EARTH UNAVAILABLE screens are gone. We have instruments now. They are reading.

There is a mountain.

Someone — and it is unclear who authorised this — acquired an ultrasonic humidifier and installed it beneath a sculpted terrain feature. The mountain glows when the humidifier runs. LED backlighting, built in. At first we thought it was a display malfunction. Then we thought it was a fire. Now we understand it is the energy source — the reason the deep space surveys flagged this sector before the colony was planted. The survey reported an anomalous thermal and optical signature at low elevation. What they detected was a small plastic mountain with a blue LED and a mist outlet.

We are mining it anyway.

Misting operations conducted: the Bulbosa received its second scheduled mist since Mark 031.

The Haworthia Superwhite received 10 millilitres of water. Measured. Careful. These things drown easily.

All systems nominal.

The display alternates screens every five Ticks — Earth side dark, World side ticking. The Mark is the date. The Ticks and Pulses are the clock. A white rule divides them, one pixel thick, drawn across the middle of the panel.

MQTT telemetry flowing: temperature 22°C, pressure 1006 hPa, humidity still settling after a rough reconnect. World time publishing to home base every sixty Ticks. The core worlds can watch the Marks accumulate.

Sleep mode is live. Every night at 2200 Earth time, every running light in camp goes dark — all twelve of them, from Pi ACT indicators to the OLED itself. The outpost breathes in silence until 0800. No blinking. No standby glow. Just the specimens, waiting in the dark.

The splitter held today. Both devices on the bus, both reading clean. One small victory in a long campaign.

The GPS module isn't here yet. When it arrives, Earth time will sync. The top half of the display will show real coordinates and a real timestamp from orbit. The colony will know exactly where it sits in the wider universe.

For now, we know where we are well enough.

MRK 022. The colony endures.

Two clocks now run on this outpost.

Earth time — the time the rest of the universe keeps. Currently unavailable. The GPS antenna and satellite dish are not yet deployed. The chronometer reads NO SATELLITE — EARTH UNAVAILABLE.

World time — the time that matters here. Running at 1:450 scale. One Mark passes for every 7.5 Earth hours. One Sol will take 450 Earth days to complete.

We are on Mark 16. The colony is sixteen local days old.

In that time the lithops have not visibly moved. They are, according to survey data, the fastest growing organisms on this world. The reports describe aggressive megaflora — spreading, consuming, reshaping terrain at alarming speed.

In Earth time, they are small succulents sitting quietly in a terracotta pot.

This is the nature of this place. Everything here moves at its own pace. You have to learn to read time differently.

Tried to bring the display screen online today.

Should've been simple. Wire the OLED parallel to the atmospheric sensor on the same I2C bus — 0x3C for the display, 0x77 for the sensor. Simple enough for any frontier electrician worth their salt.

Except nothing on the frontier is simple.

The moment that screen touched the bus, the whole line went dark. i2cdetect showed nothing. Not the display. Not the sensor. The bus locked up tighter than a cargo hold in atmo.

Spent what felt like hours chasing it. Turned out the splitter was flaky as hell. Loose connections. Wrong slot. Each time you reseated one device, the other would drop off. Maddening work.

Found out along the way: the sensor had quietly shifted address. The SDO pin, floating. The address pin drifted when the bus got disturbed. Sensor is now officially at 0x76 — chip ID 0x58 confirmed. Updated all records accordingly.

The display came to life on 5 volts. Then again on 3.3. The charge pump on these cheap modules is temperamental. But she runs.

The DSEG7 font on a 128×64 panel looks exactly like an instrument readout from a salvaged freighter. Which is exactly what this is.

Getting word out to the core worlds has been a trial.

The WiFi dongle — an RTL8188CUS, older than most of the crew — flat-out refused to negotiate with the new network stack. Something about WPA3, PMF, 802.11w. Security protocols the old girl never learned. Had to bypass the whole NetworkManager and wire her up direct: wpa_supplicant, hand-rolled config, ieee80211w disabled. Old protocol. Dirty. Works.

Now she talks. Telemetry packets flying to the relay station at 192.168.88.197. Home base can see us. Temperature. Humidity. Pressure. System load. The little details that tell you something's alive out here.

LED controller online. Home base can now toggle our running lights from the core. Which mostly means turning them off so we can sleep. Civilisation.

Three Ticks and twelve Pulses into this world's history. The specimens have not visibly moved.

In world time, it has been three and a half local days.

Made landfall today.

Five specimens acquired from the surface markets. Don't rightly know what we're dealing with yet — the locals call them lithops, which in the old tongue means "stone face." Fitting. They sit in the soil like something waiting. Not quite plant. Not quite rock. The xenobotanists back rim-side say they're harmless. The xenobotanists back rim-side have never been here.

Brought in the compute core — an old Raspberry Pi 1 Model B+, vintage hardware by any measure, but she still runs. Set her down in the dust alongside the specimens and declared this place a colony. Planted a flag, metaphorically speaking. Named her outpost-pi.

The atmospheric sensor is wired in. BME280, address 0x77. She's reading 22 degrees, 1006 hectopascals. Air is breathable. Barely.

Clock is running. Mark zero. World time begins.