Domain proposals
A candidate technology domain for the flagship facility is proposed as one Markdown file in this directory. The proposal exists so the screen in decision D12 runs before research money does.
Seven candidates were each researched to completion and each died of the same thing: a process architecture change on the same cost line beat the materials discovery, including in the two cases where the process genuinely did not yet exist. Value that reduces a loss is bounded below by zero loss, so it saturates, and the incumbent has owned that cost line for decades.
Every one of the seven was killable in under a day, and five of them in under an hour.
Start from the template, which carries every required heading and the reason each one is there.
What a proposal must contain
tests/test_domain_proposal.py fails the build if any of these headings is absent, and fails it
again if the controlled cost share is stated without a percentage or falls below 15%.
| Section | The question it answers | Decision |
|---|---|---|
| Value function | Can the customer buy an article today that does the same job? If so the value saturates. State the form: threshold, or multiplicative into an unsold quantity. | D12, test 0 |
| Physics-limit sufficiency | The property at its ceiling, divided by what parity with the incumbent's best available architecture requires. R ≥ 3. | D12, test 2 |
| Architecture ratio | The best zero-discovery flowsheet change, divided by the materials axis at its ceiling. A < 0.3. This check has caught seven of seven. | D12, test 3 |
| Controlled cost share | What percentage of the payer's cost does the discoverable property govern? Show the arithmetic. | D12 |
| Process maturity | Does a mature incumbent process already exist that this would merely improve? | D6 |
| Annual tonnage | World annual tonnage of the addressable output, with a source. | D6 |
| Attribution distance | How many transformation layers sit between the advance and the payer? | D6 |
| Computational regime | SOLVED, PARTIAL or BLIND. Only PARTIAL qualifies. | D6 |
| Measurement identity | Is the number the literature reports the number that sets cost? | D12 |
| Fast screen predicts slow truth | Does the cheap measurement genuinely predict the expensive one, and with which sign in the variable that must change at scale? | D6, D13 |
| Capital intensity | Discovery value per unit of output against capital charge per unit of output. | D11 |
The check that closes the convention's hole
A rule of the form "put it in this directory" is escapable by writing it in a different directory. So a third check reads decision D6 directly: once that entry stops reading as open, it must link to a proposal file that exists here. A domain choice cannot enter the decision log without a document the screen has already been applied to.
Why the floor is 15%
A property governing under roughly a sixth of the payer's cost cannot carry a venture outcome, however good the science is and however large the market. Market size does not rescue it: the share is multiplicative with the market, so the arithmetic is the same at every scale.
The floor is a screen rather than a verdict. A proposal that fails it is not necessarily wrong about the physics — it is wrong about who would pay for the physics, which is the cheaper thing to discover first.