Skip to content
PithomLabsPublic

About

No description, website, or topics provided.

Resources

Stars

0 stars

Watchers

0 watching

Forks

Latest commit

 

History

26 Commits

Folders and files

NameName
Last commit message
Last commit date
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 

Repository files navigation

PhysMath — theory-neutral mathematical representation

github.com/PithomLabs/phys-math is the pure mathematical common denominator of the Phys architecture. It represents mathematical structures and executes a closed set of explicitly specified operations — without knowing general relativity, quantum mechanics, Bohmian mechanics, or any other physical interpretation.

Every exported API must be explainable without naming a physical theory. purity_test.go and the TestNoPhysicalVocabulary / TestImportDiscipline gates enforce that boundary.

What it owns

  • Structured expressions: symbols (ephemeral, bound-parameter, constant), rationals, arithmetic nodes, powers, function calls, relations (ast.go).
  • Functions, parameters, and identity: declaring functions with explicit parameter vocabularies, so differentiation targets are never inferred from display names (identity.go).
  • Tensors, geometry, and coordinate structures, with rank, slot, variance, and space discipline.
  • Derivations, theorems, equations, conditions, and obligations as durable, content-addressed artifacts (families.go, store.go).
  • Nine closed executable rules — and no others: EXPR-CANON-001, EXPR-SIMPLIFY-001, EXPR-SUBSTITUTE-001, RAT-NORMALIZE-001, DIFFERENTIATE-001, TOTAL-DERIVATIVE-001, EXPR-COMBINE-001, SELECT-BRANCH-001, EXPR-APPLY-001 (rules.go).
  • Replay (replay.go): re-executing a recorded derivation from its declared premises, inputs, parameters, and conditions.

What it explicitly does not own

Physical meaning of any kind: no theories, assumptions, conventions, interpretations, or trust semantics. It also does not provide unlimited symbolic mathematics — there is no general distributivity, term collection, or rational-function normalization beyond the nine rules, and an equivalence the engine cannot derive is not engine-proved (it may still be checked by independent, hand-derived evidence in the consuming fixture).

How it fits into Phys

phys-lib/<framework>   physical meaning (assumptions, equations, theorems)
        ↓ references pure mathematics
phys-math              this repository: representation + bounded execution
        ↓ encodes through
phys-artifact          canonical bytes + identity

PhysArtifact supplies identity and encoding; PhysLib adds framework semantics on top of these representations; PhysTrace coordinates paper-derivation workflows. Related repositories: phys-artifact, phys-lib.

How do I start?

go build ./...

Read doc.go, then AGENTS.md (operating notes and the test-shard protocol), ast.go, rules.go, and replay.go. Fast feedback loop (registry, purity, codec — seconds):

go test -count=1 -run 'TestRegistryCompleteness|TestNoPhysicalVocabulary|TestImportDiscipline' .

The full suite is sharded by cost; see AGENTS.md for the exact shard commands and timeouts before running it.

How does derivation formalization work?

A derivation is an ordered list of steps, each naming its producer, rule, exact input references, conditions, parameters, and output reference. Intermediate results are minted artifacts, so provenance is a chain of references rather than prose. Replay re-executes each step and fails closed on missing, wrong-family, or corrupted evidence. Structural validation checks grammar without executing; replay checks execution without judging physical meaning.

Where are the rules for representing derivations?

In the normative Scientific Derivation Specification Guidelines, which govern faithful, auditable scientific derivation formalization. This README is orientation only and does not replace them. A replayed derivation establishes that the recorded computation reproduces from its premises — not that a theorem is true, and not that any physical claim holds.

License

MIT

About

No description, website, or topics provided.

Resources

Stars

0 stars

Watchers

0 watching

Forks

Releases

Packages

Contributors

Languages