ADR 002: Signal Path Strategy (Alphabet Index vs. Raw Chars)

Status

Accepted

Context

The internal signal path of the Enigma Machine involves passing a character through multiple rotors and a reflector. There are two primary ways to represent this signal:

1. Raw Characters: Passing char values (e.g., 'A', 'B') and performing character arithmetic.
2. Alphabet Indices: Passing AlphabetIndex values representing the index in the alphabet (e.g., 0 to TRANSFORMER_SIZE - 1).

Decision

We will use Alphabet Indices (AlphabetIndex values, 0 to TRANSFORMER_SIZE - 1) for all internal signal transformations within the Transformer, RotorBox, and PlugBoard classes.

1. Semantic Consistency: Transformations are mathematical mappings within a set. Using indices makes this mapping explicit.
2. Performance: Avoids repeated conversions between ASCII/UTF-8 and numeric indices during the hot path of encryption.
3. Flexibility: Easily supports non-standard alphabet sizes (e.g., 29 characters) by simply changing the TRANSFORMER_SIZE constant and updating the index range.
4. Boundary Responsibility: The EnigmaMachine::keyTransform (or a higher-level UI wrapper) is responsible for converting input char to int and back to char for the final output.

Consequences

• Validation: All internal components must enforce that inputs are within the valid [0, TRANSFORMER_SIZE-1] range.
• Type Safety: While int is used currently, future refactoring should consider strong typing (e.g., an AlphabetIndex class) to prevent accidental mixing of indices and raw values.