implementation-structure-design

To decide on code reuse versus new module creation before coding, enforce a reuse-first placement decision that evaluates existing functions, classes, or services for suitability, rejecting speculative abstractions and only introducing new structures when reuse is proven impossible.

Choosing between object modeling and functions for code structure requires evaluating identity, invariants, encapsulation needs, and lifecycle boundaries to decide between records, services, adapters, strategies, and classes, enforcing single-responsibility decisions for the new implementation.

The best way to enforce directory boundaries and file placement for clean architecture is to group new files by cohesion and real layer capabilities, avoiding dumping-ground structures while preserving architectural dependency direction across backend and frontend capabilities.

To prove your code structure works correctly, place tests at observable behavior boundaries next to the capability being tested, ensuring the chosen structure is provably usable and that internal modeling decisions support clear, observable interactions.

You should not use speculative abstractions during code refactoring when extracting or moving code, because forcing an explicit implementation-structure decision rejects premature generalizations and limits shared utilities to pure technical helpers to maintain high cohesion.

Yes, you can apply implementation structure design to both frontend and backend capabilities, as the process evaluates abstraction placement and responsibility boundaries for new additions, moves, extractions, and reorganizations across the entire stack before code is written.