UTF-4088 Large Sampling Map
Purpose
This map is the next feasibility stage after the 1M raw-capacity experiment and the 5M interpretability harness.
Default scale: 100,000,000 deterministic 96-bit samples.
The program records each glyph's topology and whether it survives the current engineering candidate filter.
Candidate filter
A candidate must be:
- 4-neighbor connected;
- 8–40 black pixels;
- 8–90 horizontal/vertical transitions;
- at least one graph edge.
These criteria are deliberately computational. They are not claims that the resulting bitmap has human linguistic meaning.
Outputs
utf4088-large-sampling-map.csv — per-sample measurements.
utf4088-large-sampling-map-summary.csv — density distribution and candidate density distribution.
Run
large_sampling_map 100000000 utf4088-large-sampling-map.csv
The default seed is fixed at 0x40880812 for reproducibility.
Feasibility calculation
Let r be the observed candidate rate. The estimated population is:
N_est = r * 2^96
The target is four billion:
N_target = 4,000,000,000
The observed rate is sufficient if:
r >= 4,000,000,000 / 2^96 ≈ 5.0487e-20.
Thus even an extraordinarily small candidate rate can theoretically support the target. The practical purpose of the large sample is therefore not to prove raw capacity, which is already established, but to estimate how restrictive the graph/visual filters are and to locate useful regions of the glyph space.
Map interpretation
The map is a distribution map, not an assignment of meanings. Historical and language priors are applied only after the candidate population has passed the geometric/topological stage.
The intended progression remains:
96-bit space -> graph-valid map -> visual candidate map -> historical/language priors -> 16,606 curated layer -> 4B experimental namespace
Execution note
The GitHub repository stores the executable source and map format. A GitHub contents operation does not execute native C++ code, so numerical output must be generated by a local/CI build runner. The repository should therefore retain the exact seed, filter, command, and expected report schema so that a CI runner can publish the generated CSV and summary without changing the experiment.