The Apollo Lunar Module had a checklist for cabin depressurization, for surface egress, for sample stowage. It had no checklist for the regolith the crew tracked back inside — electrostatically charged, fifty times finer than a human hair, sharp like glass. All twelve moonwalkers reported irritation anyway. Harrison Schmitt named it lunar hay fever on Apollo 17. The cabin smelled like burnt gunpowder for hours after the EVAs.

Draw the mission as a dependency graph and the failure mode becomes obvious in retrospect. The architecture had regolith as a surface input — sample science, lens fouling, suit abrasion. It did not have regolith as a respiratory input. The pulmonary path was an unlabelled edge. Twelve flights, twelve cases, no one drew the line.

The 2018 ESA write-up that resurfaced on Hacker News this week is not new science — it is an admission that the dependency graph still needs to be drawn. Kim Prisk, a UCSD pulmonary physiologist with twenty years on human spaceflight, is one of twelve scientists working through ESA’s lunar-dust toxicity programme. The simulant comes from a German volcanic region, ground from raw material that — and this is the irritating part — has its sharp edges removed during prep, because true lunar fines are too rare and too dangerous to grind freely. Erin Tranfield, the biologist running that side of the work, called it "a special kind of dust." That phrase is doing a lot of work.

←TODAY: ESA’s Airway Monitoring experiment is still running on ISS in 2026 to characterise lung response in reduced gravity, and Artemis crews are months from putting boots back on the regolith.
→3012: Closed-loop particulate control becomes a first-class subsystem of any habitat with airlock and crew — Moon or Mars — equal in status to oxygen and water management.
Fulcrum: The Apollo crews’ allergies were not a medical anomaly; they were the third single-point-of-failure the program did not know it had. Find yours before the system tells you.

The architecture parallel is uncomfortable. Silicate inhalation is a known risk on Earth — miners’ lungs have been telling us this for two centuries — but in BIM coordination the generation of construction dust is rarely modelled. There is no clash-detection pass that says "your circular-saw cut here fills this volume of corridor with respirable silica for forty minutes." The dependency is real, the failure mode is well-characterised, and yet on most projects it lives in the SUVA handbook, not in the model.

Atelier: On PAZ-Atelier studio projects we treat fine-particulate generation as a first-class clash. When the cut list calls for more than 2 m² of concrete-saw work indoors, the air-handling plan sits in the same coordination meeting as the structural sequencing — not later, not separately. The Apollo lesson scales down: if a hazard is not in the model, the model is lying.

Hack: This Hack teaches you to compute how long a fine particle stays airborne, so the number lives in your sequencing notes the next time you spec an indoor cut. Medium: Python. Domain: physics — Stokes’ law for settling velocity in still air.

# Stokes settling velocity, 1 micron silica in still air at 20 C
rho_p, rho_air, mu, g, r = 2600, 1.2, 1.8e-5, 9.81, 0.5e-6
v = (2/9) * g * r**2 * (rho_p - rho_air) / mu
print(f"{v*1000:.4f} mm/s  ->  {1/v/60:.1f} min to fall 1 m")

Run it. A 1-micron silica grain falls a metre in roughly four hours. The cut you finished at 11:00 is still in the breathing zone at lunch. On Earth we have OSHA limits and SUVA inspectors; on the Moon there is no regulator, the gravity is one-sixth, and the hang-time multiplies by six.

So: walk your next project’s risk register the way Apollo’s planners did not. Ask which inputs cross into the breathing zone, which surfaces feed which lungs, which "known" hazards have no edge in your model. The Moon program rediscovered the dependency twelve times. Yours does not have to.

Sources & Further Reading

• Primary: ESA — The toxic side of the Moon