# Changelog

## Score

- CAI 66 → 64 (-2.4)
- 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 48 → 46 (-2.9)
- Readiness 71 → 71 (+0.0)
- Security 85 → 78 (-7.4)
- Event Sourcing 100 → 100 (+0.0)

## Resolved (3)

- Duplicate creation methods across `ActorSystem` and `ActorRefFactory`. `ActorSystem` implements `ActorRefFactory` (implied by context) but also exposes its own `sys_actor_of*` methods. This creates confusion: should users call `system.actor_of(...)` or `system.sys_actor_of(...)`? The `sys_` prefix suggests system-level actors, but the duplication of the entire API surface is redundant.
- Inconsistent messaging API across actor reference types. `BasicActorRef` has `try_tell` and `try_tell_any`, while `ActorRef` has `send_msg`. `ActorReference` has `sys_tell`. There is no unified `tell` or `send` method across the hierarchy. `try_tell` implies failure handling, while `send_msg` might panic or block, but the naming doesn't clearly communicate this difference. Additionally, `Tell` trait has a `tell` method, but it's not clear if it's implemented by all actor refs.
- Redundant factory interfaces with parallel method sets. `ActorRefFactory` and `TmpActorRefFactory` expose nearly identical method signatures (`*_of`, `*_of_props`, `*_of_args`) but are separated by a naming convention (`tmp_` prefix) rather than a clear semantic distinction in the type hierarchy or method names. This forces users to choose between two interfaces for conceptually similar operations (creating actors with/without names/args).
