Blocker-first research program
Blocker Ledger
One objective, ten blockers. This page is permanent: each blocker keeps its status, the measurement that justifies it, and the experiments that touch it. A blocker is never marked resolved physically — the best available status inside a toy model is RESOLVED IN TOY MODEL.
Core objective · one target for the whole lab
computing gate…Find a nontrivial transition law under which relational location changes substantially while identity and required invariants are preserved along an allowed path, without violating the model's locality / propagation rules.
Every experiment on this page exists to close a named blocker. If a proposed experiment does not address an open blocker or test a proposed missing rule, it stays in the ideas backlog.
Primary blockers
Running the constraint stack to compute live blocker statuses…
Ideas backlog
Discipline rule: an experiment enters the main roadmap only when it addresses a named open blocker or tests a proposed missing rule. Everything else waits here until it can be framed that way.
Redundant encoding search
Interesting, but does not address a named open blocker until it is framed as a test of B1 circularity.
Emergent conservation search
Only promotable once tied to B3 with a derivation, rather than a search for coincidental constants.
Higher-order transition operators
No named blocker yet. Would need to show it changes the identity-tolerance boundary of B2.
Stochastic corridors and annealing
Would relax pathwise validity statistically; parked until B6 has a deterministic answer.
Alternative observables on the relational graph
Instrumentation, not a blocker test. Stays in the backlog.