Contribute

Contribute

Phosphoros is a plan anyone can check, and it gets better each time someone shows part of it is wrong. The most useful contributions make it more honest: a number that is off, a source that says something else, a physics check that should fail and does not, or an invention that does a job better within the laws of physics.

Every number is marked:Physics: Measured, or established physicsPrecedent: Demonstrated in the real world, citedInvented: A breakthrough the plan assumesDesign choice: A size, count or date the plan choosesSelect any number to see how it is computed.

How to propose a change

Every number lives in data/ and flows through the model in calc/ to every page, chart and table. So a change to a number is a change to the data or the model, never to a page. Open an issue to discuss it, or send a pull request that makes it.

A number that was published is never quietly changed. A pull request that changes one also adds an entry to data/corrections.yaml saying what changed and why, and it shows up in the corrections log. An open problem gets an issue and a failing check in the model, so it stays visible until someone solves it.

Three kinds of issue

Challenge a number. Open a "challenge a number" issue. Name the number's id from the ledger, say what you think it should be, and show why: a source, a calculation, or a flaw in the formula. Every row of the ledger links to the line of code that computes it.

Propose an invention. Open a "propose an invention" issue. Say what job it does, what it replaces, what is invented and what already exists, and which numbers it changes. An invention may assume engineering we do not have. It may never break conservation of energy or the second law of thermodynamics, and the proposal should show it does not.

Add a source. Open an "add a source" issue, or edit data/sources.yaml directly. Say which claim it supports or contradicts. Primary sources beat reports of them: version 3 replaced one news item that had a launch date backwards.

Rerun the numbers

You need Node 24 and pnpm 10.

pnpm install
pnpm calc                                   # every derived number, then the physics checks
pnpm calc --only veil-distance              # one number: value, formula, inputs, where it is defined
pnpm calc --set foundry-doubling-days=70    # change an assumption and see what follows
pnpm calc --claims                          # every number in the version 2 brief against the model
pnpm test                                   # the model's invariants
pnpm docs                                   # rewrite every number in docs/ from the model
pnpm lint                                   # formatting, types, stale docs, em dashes, hand-typed numbers

The model reads measured facts from data/constants.yaml and the plan's choices from data/assumptions.yaml. Each module in calc/src/model/ computes one system and records the inputs every number read, which is how the ledger can show each formula. The Explorer runs the same model in your browser, with a slider for every assumption that has a range.

Open questions

These are the places where the plan is weakest, or least sure of itself. Each is a good first contribution.

  1. Lowering the ice. The Braking Ring catches the ice, but lowering it on tethers needs times the lift its rotor has to spare. A solution must bring to the surface by the deadline without exceeding the Ring's lift, or show how the cooling absorbs the of extra heat if the Bricks are simply dropped. The options so far are on Risks.
  2. How much water Ceres has. The model takes Ceres to be water by mass, the top of a range estimated before the Dawn mission (McCord and Sotin 2005). Post-Dawn interior models are the next source to check. The answer sets how much of Ceres's water the plan uses, now .
  3. Basalt's thermal conductivity. The model uses , the upper of two published values for dense basalt (Robertson 1988). It sets the heat the crust feeds into the air during cooling and the the cryo-continent floors must shed at first. A better value, and how it changes at Venus's surface temperature, would tighten both.
  4. Electroluminescent cooling at high power. An LED has emitted more light than the electricity it drew (Santhanam 2012), and thermophotonic transfer has been measured in laboratory structures (Casado 2019), but only at tiny powers. The plan needs , and the model charges the pump work at the Carnot minimum. What limits a real tile, and how close to Carnot it can get, are the most important open numbers in the plan. A physics argument that it cannot work would be just as valuable.
  5. How sensitive the cloud-life survey can be. The gate falls on . What can balloons and droplet sample return flown from actually detect, and what would a negative result rule out? The gate is only as honest as that answer (Research Agenda).

Pull requests

Run pnpm lint && pnpm test first; both must pass. In anything you write: no em dashes anywhere, no number typed by hand into a document (link it to the ledger instead), and every invention marked as one. The full conventions, including how to place numbers, figures and generated tables in a document, are in CONTRIBUTING.md.

Code is MIT; text, data and figures are CC BY 4.0. By contributing you agree to license your work the same way.