V3

01

Why choose the lens first

One direction can yield method, measurement, perspective or application contributions with different comparisons and evidence. V3 commits to the contribution before generating, reducing the disconnect between complexity and value.

02

Three levers supply exploration material

Kernel libraries provide foundational structure, anchor libraries provide transferable shapes and refutation checks prior coverage. Mechanism-versus-combination checks ask whether coupling adds value before drafting inputs, outputs, risks and a decisive test.

03

From self-evaluation calibration to evidence discipline

V3.1 handles originality anchors and hybrid baselines; V3.2 checks theorem applicability and weakness responses; V3.3 examines universal and causal claims; V3.4 sharpens measurement, limits promises and tracks refinement gains. The refinements narrow the gap between a proposal and evidence already obtained.

04

Public tools and historical snapshots

The repository named v3 is the assessment tool and already contains later assessment protocol content; the public generator is an early three-lever system. The local Lens-First V3.x retains that division of roles, without establishing identical code or publication of the same version.

05

What carries into V4

Lenses, strongest priors, testable questions and evidence discipline become V4 search constraints. V4 adds state, tree and ledger to make these objects persistent and resumable, while V3 remains centered on Markdown protocols and reference libraries.

How this series evolves

1. V3.x / V3.4

   V3.1 · Calibrate self-evaluation and combination risk

   Adds originality anchoring, cross-class hybrid baselines and audits of introduced weaknesses and stability evidence.

2. V3.x / V3.4

   V3.2 · Audit theorem applicability and change contribution lens

   Checks theorem assumptions, requires specific testable responses to weaknesses and suggests a new lens when the current framing saturates.

3. V3.x / V3.4

   V3.3 · Constrain empirical correction and universal claims

   Adds empirical-correction scoring anchors and requires suitable evidence and identifiability for universal and causal-decomposition claims.

4. V3.x / V3.4

   V3.4 · Separate measurement plans from observed evidence

   Adds six measurement evaluation axes, limits future commitments as evidence and monitors diminishing returns across refinement rounds.

Versions in this series

What carries forward

The contribution determines comparisons and evidence.

Kernels, anchors and refutation jointly develop candidates.

Planned work and available evidence remain separate.

Protocol revisions within the archive

Four protocol revisions retained in the same V3.x archive.

1. V3.1

   Calibrate self-evaluation and added complexity

   V3.1 responds to generation–evaluation feedback on long-tail knowledge recall: self-evaluation was high, and added components introduced stability, comparison and reproducibility risks. The patch separates complexity from originality and reassesses the actual contribution instead of continuing to raise the previous self-score.

2. V3.2

   Audit theorem conditions and permit reframing

   V3.2 addresses formal compliance left by the previous rules: a stability theorem can be cited without checking its assumptions, and a weakness response can look complete without a concrete test. Repeated refinement within one method framing can also reach structural limits, making another kernel change an unhelpful next step.

3. V3.3

   Set evidence criteria for empirical correction

   V3.3 follows a shift from method contribution to empirical correction. Earlier calibration mainly served methods, while the new framing requires belief update, experimental design, confound control and reproducibility. Reusing the old criteria can inflate confidence. Lens switching must change evaluation criteria, not merely the title.

4. V3.4

   Separate measurement evidence from promises

   V3.4 turns to measurement and benchmark research, refining its anchors through proposal feedback. It asks whether measures match constructs, coverage spans domains, baselines are adequate, reliability has data, errors are diagnosable and tools have community utility. Sub-suite counts and future replication plans cannot substitute for those answers.

Early public source

Two public tool lines are verified: assessing existing ideas and generating candidates from a direction. The public v3 repository belongs to assessment; a matching public commit for local Lens-First V3.x was not found.

The local Lens-First V3.x introduction draws on its archive and revision records; the two public tool lines separately supply early generation and assessment context.