Maintenance / patch-like
Presentation, metadata, spelling, accessibility or implementation repair that does not change the decision logic.
Trust and governance · Framework history
A version number is useful only when it tells you whether the decision method changed.
This history is designed to separate editorial maintenance from methodology evolution. A new number should communicate change meaning, affected surfaces and decision impact—not simply make the page look governed.
Method lineageDecision impactNo version theatre
Three change meanings
SERPIFAI uses patch-like, minor-like and major-like language as a communication aid. It does not claim strict Semantic Versioning compliance for editorial methodologies.
CURRENT METHODStable decision contract
Material compatible evolution
A clarified or expanded method that preserves the central decision contract but changes how it is applied or explained in a material way.
Breaking / major change
Core classification, required evidence, gate, output meaning or safe-use boundary changes enough that prior outputs may need reinterpretation.
Materiality matrix
| Change | Class | Decision consequence |
|---|---|---|
| Typography, layout, responsive repair | Maintenance | No change to decision method. |
| Source refresh with same supported conclusion | Usually maintenance | Record source review where material; do not imply method change. |
| New explanatory example | Usually maintenance | Version only if it changes the method contract or safe use. |
| Weight, threshold or gate change | Material methodology change | New method identity and parity/fixture review. |
| New classification or changed output meaning | Material methodology change | Assess migration and prior-result comparability. |
| Removal of a method because it is unsafe/obsolete | Retirement | Preserve a clear superseded/retired record. |
What a real version record contains
Method name
The stable public identity of the decision method.
Effective version
The method identity that produced a given output.
Change summary
What changed in decision-relevant terms, not a generic 'updated'.
Affected surfaces
Article/page, web tool, workbook, PDF, schema, downloads or other product components.
Decision impact
Whether old outputs remain comparable, need reinterpretation or should be retired.
Evidence + tests
Which fixtures, parity checks and direct tests were rerun after the change.
Current public method entry
The Competitor Fit Challenge currently publishes Methodology 1.1. Version 1.0 remains part of the historical lineage; it is not the current public contract.
Earlier public baselineMethod v1.0Superseded by the accepted 1.1 contract
→
Current public contractMethod v1.1Accepted and browser-reviewed
Public terminology clarified
Relationship Fit is the canonical public term. Framework Fit remains the tool-facing label for the same deterministic six-lens score.
Evidence stays separate
Evidence Readiness remains a separate decision layer. It is not multiplied into the fit score, and weak evidence can produce MORE EVIDENCE REQUIRED.
Comparability preserved
The 1.1 reconciliation did not authorize silent confidence arithmetic or a hidden replacement method. Future weight, gate, output-meaning or safe-use changes still require a new accepted method record and retest.
LINEAGE RULE
Superseded does not mean erased
Older method identities remain part of the record so a prior output can be interpreted against the method that produced it. A newer accepted contract becomes current only through an explicit change record; it does not silently replace history.
Two ways version histories fail
Questions about versions
Does every page edit need a new method version?
No. Maintenance that does not change decision logic should not create a method-version event.
Why mention Semantic Versioning if this is not software API versioning?
Because SemVer demonstrates a useful principle: the number should communicate the meaning of change. SERPIFAI uses that principle as an analogy, not as a claim of strict compliance.
What happens when a method changes materially?
Create a new identity, record the change and decision impact, update affected surfaces and rerun relevant parity/direct tests.
Can old outputs remain valid?
Sometimes. The version record should state whether the change affects comparability or safe use rather than assuming every older result is invalid.
What is the current Competitor Fit Challenge methodology version?
Methodology 1.1 is the current accepted public contract. Version 1.0 is retained as an earlier lineage entry so older outputs can be interpreted against the method identity that produced them.
Track three identities when a decision product evolves
This prevents a major interface redesign from being mistaken for a new analytical method—or a material rule change from hiding inside a minor product release.
Interface and workflow
Navigation, responsive composition, progressive disclosure, evidence visualization and interaction behavior.
Decision semantics
Dimensions, weights, gates, precedence, controlled states, output definitions and refusal conditions.
Current support
Source observations, first-party evidence, verification date, known contradictions and reopening triggers.
Example: moving from a dense graph to Guided Answer → Proof → Next can be a substantial product improvement while the method remains stable. Changing a critical gate is different: that can alter which decision the same evidence supports.
Evidence and source roles
Semantic Versioning informs the principle that versions should communicate change meaning; COPE informs visible correction/public-record discipline. SERPIFAI's methodology version model is its own governance practice.
See how changes are classified
MONITORING OR REFERENCE CONCLUSION
Open correction policy →Corrections and Methodology Change
The correction policy explains when maintenance becomes a factual or method-level change.