Long Count
Tala woke because the ship had become too quiet. The drive was idle, and the traffic receiver had stopped murmuring through the bulkhead beside her berth. She pulled herself along the ceiling rail to the galley, where Sana had strapped the gray book to the chart table.
The last traffic return had vanished during her sleep, twelve hours after the Archive fell behind them. Vela crossed a region where aid required weeks and the ship's waste heat outshone every neighboring object.
Passage watches replaced station time. Mara divided sleep into short pieces. Osei listened through meals for changes in pump tone. Every morning Sana opened a new account page, even when vacuum had supplied nothing to enter.
On a small ship, privacy belonged to the watch schedule. Tala learned the creak outside Sana's berth and the difference between Mara's contented silence and her dangerous one. Osei disassembled healthy devices whenever Sable's distant repairs occupied him. Questions about reassembly only extended the process.
Tala kept Venn's folded reference inside her private locker and read it when the ship was quiet. The page warned that she would halt apparently healthy machinery when its model omitted something physical. Some nights that sounded like judgment. Some nights it sounded like a machine fault wearing her name. She had joined Vela without learning which version was true.
Meals attached themselves magnetically to the chart table. The galley could provide chairs for the crew or screens for the watch. Mara defended the arrangement as design, Osei as scheduling. Sana had opened an inquiry into the missing chair.
The gray book traveled with the watch.
Tala opened the route map above the table. The Giant occupied its center as a striped disk. Catenary's wheels clustered deep inside the magnetic field to the right of Vela's track. Sable moved far to sunward on an irregular captured orbit, close to the boundary where stellar wind first pressed against the Giant's field.
She advanced the weather layer. The charged front crossed Sable's path first. Only later did the model carry it inward to Catenary. Sable's independent clocks and instruments therefore felt a change while the traffic lanes still had time to act.
The Second Navigation had grown around that geometry. Catenary supplied the traffic network and an optical reference. Sable supplied a comparison made on the weatherward side. No vote at either port could manufacture the other's observation.
Early surveys had processed Sable's darkness as an absence in their data. The instrument orbiting that moon greeted Vela with thirty-two binary observations and no explanation.
0 1 1 0 0 1 0 0
1 0 1 1 0 0 1 0
0 0 1 0 1 1 0 1
1 0 0 1 0 1 0 0
Mara moved the digits onto her navigation pane. Old Reach pilots had pulled pulsar fixes through charged dust, and binary ephemerides still hung above some Thorn bunks. Ordered lights invited her to find a road. She followed eight with one finger, then withdrew her hand.
Osei searched for a header. Sana traced the transmission and found that each digit came from a fresh preparation at the same remote source. The sequence carried no route. It was the instrument's greeting, a small test of what the receiving crew would invent before the real data arrived. Sable's camps had inherited too many charter reports whose confidence grew with transmission distance; the cooperative now tested the interpreter as carefully as the link.
Experimental procedure decided what the sequence meant. Repeated readings of one evolving system would form a history. Here the instrument made a fresh preparation for every reading, producing thirty-two trials.
Tala called each complete preparation and procedure a trial. Observation ended it with one outcome. This interface admitted 0 or 1; each trial supplied one of them.
Mara isolated the first 0. "What have we learned about the source?"
"This preparation produced zero under this procedure."
Mara selected the following digit. "And this?"
"A separate trial produced one."
"The Reach has made you less informative."
Tala shared the disappointment. Sana's approval confirmed that responsibility had caused it.
Catenary surrounded every number with neighboring systems and people available to contradict it. Across the Reach, invention could arrive hours ahead of correction.
A single outcome could answer only its own trial. Sana preserved all thirty-two in order and named the finite collection a sample.
With the individual outcomes preserved, Tala counted.
outcome 0: 18
outcome 1: 14
total: 32
Counting left the observations unchanged while discarding their order. Sana retained the sequence beside its totals, following the Archive practice of displaying what a summary had consumed.
Mara's Archive bird stood upright during thrust and searched slowly for down whenever the drive slept.
The Common Book gave every counted trial one visible mark.
0 | ################## 18
1 | ############## 14
Those bars formed a histogram of observed counts. They depicted neither amplitude nor confidence nor a wave in the chamber. Any permutation of the sequence would preserve the histogram while changing the ordered sample.
Osei converted the counts by dividing each one by thirty-two.
| Outcome | Count | Frequency |
|---|---|---|
0 | 18 | 18 / 32 = 0.5625 |
1 | 14 | 14 / 32 = 0.4375 |
Each resulting frequency described the finite sample. A later packet finally delivered the source specification and numbers of another kind.
model probability of 0: 0.5
model probability of 1: 0.5
The probability model assigned weights independently of this particular sample. Equal model probabilities did not require equal counts in a finite block. Frequencies away from 0.5 could agree with the model, while a perfect split could occur without proving it.
The first calibration deadline drew closer. Mara enlarged the bars, looking for a count she could place on a schedule. "How many trials finish the question?"
Sana opened the model's stated scope. "Which question?"
Conversation stopped as quickly as a hand leaving an unsafe switch.
Additional independent trials could steady an estimate when the named preparation and model applied. Quantity could not supply an unstated assumption.
Osei prepared a simulator request for thirty-two fresh trials. Target, shot count, and seed entered the mission account before the sample existed, beyond the reach of later preference.
wheeler run manual/BinaryTrials.wbc \
--target ideal \
--shots 32 \
--seed 104729
Its ordered outcomes differed from Sable's greeting. Their counts produced the same bars.
0 | ################## 18
1 | ############## 14
Mara compared the displays. "The simulator repeated them."
Sana aligned the sequences. "It repeated the histogram."
Different observations occupied the two sequences. Identical totals occupied the histogram, whose construction had discarded order.
The recorded seed made this simulator sample reproducible. Another seed would create another reproducible sample. Neither command predicted a future physical preparation without that seeded numerical model.
Sable withheld the approach reservation until Mara acknowledged the greeting. A second physical sample required six hours and would leave the same assumptions unstated. Refusal spent fuel; invention spent trust.
"Give me the sentence I can sign," Mara said.
Tala returned to the transmitted record. "These instrument trials produced the preserved sequence and counts. The attached model assigns equal probability to both outcomes. The sample alone does not establish that model."
"Your yes requires unusual cargo space."
"It fits the evidence we possess."
Mara entered the statements separately and accepted Sable's approach. She preserved contingency fuel for contradiction. Limited evidence still permitted a limited decision.
Tala arranged the three names in the gray book.
| Object | Statement it supports |
|---|---|
| One outcome | One trial ended in 0 or 1 |
| One sample | These trials produced these counts and frequencies |
| Probability model | The stated preparation and procedure assign these outcome probabilities |
An outcome belonged to one trial. A histogram summarized a sample. The model assigned probabilities under assumptions and promised no particular next digit.
Sable's carrier tone became the only nearby voice. The next packets described two modeled routes meeting at a detector, each bearing a signed quantity that ordinary probability could not represent.
One pair reinforced while another vanished. Mara shook her head. "Observed counts cannot do that."
Artificial evening dimmed the cabin. The watch continued around the new arithmetic.
The packet and the gray book met under one heading: Contributions That Cancel.