V5

01

Why V5

V4 persisted candidate search, but strong scores still needed reliable evidence boundaries. V5 first defines independent verification: when verdicts are allowed, how missing checks are represented and why failures must persist. It then expands generation because validators cannot create missing artifacts.

02

Aurora → Eureka: question, construct, check

V5.0 appends conditional gates; V5.2 links contracts, tensions, genomes, mathematical objects, definitions and conjectures. V5.3 adapts discovery objects to Aurora and reconciles verdicts. Shared interfaces emerge, while payload fidelity and numerical witnesses still need work.

03

Spark → Forge: diagnosis must affect the next step

Spark adds actual-payload indexing, canonical equation comparisons and four small probes. Forge RC implements viability, repairs and harvest gates; Final wires them into the loop and records repair lineage, workstream states and bridge-only objects. RC and Final share 5.5 naming but have different integration.

This phase distinguishes selecting a repair operator from implementing its semantics. Origin later improves special operators that fall through in Final; coordination records likewise precede content execution.

04

Origin → Genesis: explore representations

Origin introduces D/M/P paths and first-class mappings. Genesis adds compatibility, quality and diversity filters plus finite RHS execution. Equation structure begins affecting computation, and conjecture type affects proof routes. Richer representations are paired with tighter validity requirements.

05

Logos: evidence must describe one candidate

V5.8 joins evidence by mapping ID, tracks property provenance and separates broken packages from refuted ideas. Dual-form AST access repairs serialized empty-RHS risks. The sequence ends with a basic requirement: every report artifact must identify the object it describes.

06

How to read the historical pages

Each page covers the problem, flow, modules, contracts, conceptual scenario, changes, source evidence and scope. Diagrams describe artifact paths, and reading scenarios explain architecture without claiming new runs. Dedicated call paths and later audits resolve disagreements with changelogs.

How this series evolves

1. V5.0

   V5.0 Aurora

   Makes verification a separate execution layer. Candidate artifacts start unverified, and explicit checks govern promotion, rejection, unresolved evidence and honest failure.

2. V5.2

   V5.2 Eureka

   Moves upstream from filtering false originality to organizing candidate discovery: frontier tensions drive typed operators, equation genomes, mathematical objects, definitions and conjectures before Aurora review.

3. V5.3

   V5.3 Eureka

   Retains Eureka’s discovery layer while repairing its verification interface: translates discovery objects into Aurora payloads and reconciles every output to one terminal verdict.

4. V5.4

   V5.4 Spark

   Routes harvested equation payloads into verification and extends the Equation Lab with canonical forms, a prior-equation library, small numerical probes and physics diagnostics.

5. V5.5

   V5.5 Forge RC

   Adds explicit viability and bridge firewalls before idea harvesting. Failed candidates remain legible in structured failure records with repair reasons instead of being packaged as discoveries.

6. V5.5

   V5.5 Forge Final

   Wires Forge’s standalone components into the live discovery chain: resolved units, viability records, repair traces, CoMath state and prior caps contribute to one package.

7. V5.6

   V5.6 Origin

   Expands from equation-term mutation to the choice of mathematical descriptors. Domain, math and physics search trees produce first-class mappings and structured equation ASTs.

8. V5.7

   V5.7 Genesis

   Adds compatibility, execution and quality constraints to Origin: prune mappings, select diverse elites, execute supported AST terms and route proof strategies by conjecture type.

9. V5.8

   V5.8 Logos

   Closes the ontology-to-idea evidence chain around one candidate ID. The release focuses on linkage, property provenance and replay rather than adding mappings or solvers.

Versions in this series

What carries forward

Research contracts and honest failure: define the claim and retain incomplete or refuted paths.

Typed candidates and provenance: equations, conjectures, definitions, mappings and claims need identities.

Separate generation and checking: scores cannot replace execution, priors or evidence checks.

ID-linked replayable packages: final statuses should lead back to one candidate’s artifacts.