Most learning software is a content problem wearing a software costume. This one is the other way round: a deterministic calculation engine with a syllabus wrapped around it.
The engine is TypeScript with zero runtime dependencies. No network call, no server, no hidden state. Give it the same inputs a year from now and it returns the same answer, to the last decimal.
1,292 automated cases run against every commit. Many compute the same quantity by two genuinely different methods and require the two to agree — a textbook value on one side, an independent derivation on the other.
Where an approximation is used, the lab names it and states its error. Where no answer exists, the lab says so. Neither of those is a failure mode; both are the product working.
The dry adiabatic lapse rate reads 9.76 °C per kilometre because the engine divides gravity by the specific heat of air. It is not a number somebody typed in from a textbook.
The plane-chart approximation used for coastal fixes is out by 0.3 metres at 12°N and 4.4 metres at 51°N. The laboratory prints that, next to the answer, unprompted.
Squeeze the isobars around a high tightly enough and no wind balances. The lab says that. It does not round to something plausible and move on.
Every glossary term is linked to the laboratory that simulates it, so a definition is never the end of the road. The whole platform is a progressive web app: once it has loaded, it works with no connection at all — which is the difference between studying at sea and not studying at sea.
Cosmic energy inside the ring of fire, perfect equilibrium at its centre. Energy all around the ship; stability by design. That is the syllabus in one image — which is why the name stuck.
CERN keeps a Nataraja statue outside its buildings, for related reasons.
“The still point of the turning sky.”
A cadet or a serving officer starts free and is inside a laboratory in a minute. A college talks to us first, because enrolment and faculty tools need setting up properly.