bloom-allocation

Extract this entire folder. Python 3.10 or newer is the only dependency for
the experiment. No package installation, account, API key, network, or GPU
is needed. From this folder run:

    python experiment.py

This regenerates data/*.csv, data/summary.json, and figures/*.svg. The
notebook uses the included experiment.py and can be opened in an existing
Jupyter environment; choose this folder as its working directory and Run All.
Notebook display uses IPython, which is supplied by a normal Jupyter Python
kernel. The plain Python script does not need Jupyter or IPython.

Files
  article.html: standalone article with relative figure and download links
  experiment.ipynb: runnable walkthrough, saved outputs, checks, and figures
  experiment.py: complete locally authored implementation and assertions
  data/: generated raw counts, summaries, and deterministic traces
  figures/: generated SVG plots and diagrams
  sources.json: primary sources, roles, and verification date
  reproducibility.zip: code, notebook, data, figures, sources, and this README

The code writes only its output data/ and figures/ directories. To retain
the shipped results, copy the whole folder before rerunning. The fixed seeds
make the Bloom counts reproducible. The quorum enumeration is deterministic.
Python version in summary.json is environment metadata and can differ.
