# Changelog

## Score

- CAI 47 → 48 (+0.9)
- Rubric changed (rubric-2026.10.5 → rubric-2026.11.2) — scores are not directly comparable.

## Lenses

- Code Health 98 → 98 (+0.0)
- Architecture 99 → 99 (+0.0)
- Maturity 45 → 55 (+10.1)
- Readiness 42 → 42 (+0.0)
- Security 43 → 43 (+0.0)
- Domain Modelling 100 → 100 (+0.0)
- Accessibility 48 → 48 (+0.0)
- Performance 91 → 91 (+0.0)

## Resolved (3)

- Naming inconsistency between public API and internal implementation. The public BTree uses 'search' while the internal Node also uses 'search'. While this might be intentional, 'search' is often ambiguous in key-value stores (does it return the value or a boolean existence check?). More critically, if BTree delegates to Node, the naming is consistent, but the return type of BTree.search is V (the value), which implies it panics or returns a default on missing keys, whereas 'search' often implies an Option/Result return. However, the primary inconsistency is between BTree.insert and Node.insert signatures.
- Signature mismatch between public API and internal method. BTree.insert does not take a 'degree' parameter, but Node.insert does. This suggests that BTree.insert must internally pass the degree, but the public API hides this implementation detail (which is good), yet the internal Node API exposes it. This is not a direct inconsistency in the public surface, but it indicates a lack of encapsulation in the Node type's public-facing methods if Node is exposed or if the API is meant to be uniform.
- Vague and inconsistent naming in utility module. 'optimize_memory_usage' is a high-level goal, not an operation. 'reuse_buffers' is an implementation detail. 'handle_large_data' is too generic. These functions lack clear, distinct purposes in their names, making it hard to understand what each does without reading the body. They may overlap in functionality (e.g., optimizing memory might involve reusing buffers).
