A minimal pseudocode language for making physics reasoning inspectable.
Humans and AI agents use it to communicate hypotheses, derivations, and paper reasoning while keeping assumptions and unresolved gaps visible on the page — so speculation is never mistaken for established fact.
- Hypothesis exploration — write down a conjecture, its premises, and what would have to hold, then examine the consequences adversarially.
- Inspecting derivations — a readable counterpart to the essential reasoning behind Phys Go derivations, for human review before implementation.
- Paper overviews — a compact map of a paper's assumptions, reasoning, and open questions, in a form that separates what the source shows from what remains missing.
given x_positive: x is a positive real number
define y := x + 1
derive y_positive: y > 0 from [x_positive, y] | adding one to a positive real gives a positive real
? whether this premise describes a particular physical system is not established; that requires separate justification or evidence
The derivation follows mathematically from its premise — but writing it down does not establish that the premise describes any physical system.
Writing an equation or derivation in Phys Pseudo does not prove it
correct or establish that a model describes Nature. The language records
reasoning; it carries no verdicts. (Formally: representation is not
endorsement — PSEUDO.md §7.)
PSEUDO.md— the canonical language specification.MRC_REVIEW_PROTOCOL.md— how reasoning written in it is adversarially reviewed.conformance/README.md— example fixtures, including deliberately flawed ones, and how to read them.
Implementation handoffs: derivations intended for implementation must
additionally follow
conformance/DERIVATION_SPEC_GUIDELINES.md.
Working in this repository as an agent? See AGENTS.md.
MIT