A02 reference (human + AI, not part of the comparison)
A human wrote it with AI help for A02 before this experiment, so it is never ranked.
Figure 23.1
Class diagram of the A02 reference (human + AI, not part of the comparison)
10 classes and 37 links. Tag: recomputed.
Text version: Class diagram of the A02 reference (human + AI, not part of the comparison)
- Package kalah: Board, BoardPrinter, Game, House, Kalah, Pit, Player, Sower, Store, Turn
- Board uses BoardPrinter
- Board uses House
- Board holds House
- Board uses Player
- Player uses Board
- Board uses Sower
- Board uses Store
- Board holds Store
- BoardPrinter uses House
- BoardPrinter holds House
- BoardPrinter uses Player
- BoardPrinter uses Store
- BoardPrinter holds Store
- Game uses Board
- Game holds Board
- Game uses Player
- Game holds Player
- Game uses Store
- Game uses Turn
- Game holds Turn
- House uses Pit
- House extends Pit
- House uses Player
- Kalah uses Game
- Pit uses Player
- Pit holds Player
- Sower uses House
- Sower uses Pit
- Sower holds Pit
- Sower uses Player
- Sower uses Store
- Store uses Pit
- Store extends Pit
- Store uses Player
- Turn uses Board
- Turn holds Board
- Turn uses Player
| Functional mark /15 | not measured |
|---|---|
| Hidden pass /76 | not measured |
| Unseen pass /69 | not measured |
| Strict PMD compliance | pass |
| Mutation score | 154/165, 0 not run (the reference's own TestKalah tests from the hidden jar) |
| Classes | 10 |
| Simple Constructs /5 | 4 |
| Minimal Duplication /5 | 4 |
| Modular Structure /5 | 4 |
| Problem Alignment (tentative) /5 | 4 |
| Weighted tokens | not measured |
| Bench median ms | 500 |
Reference
The A02 reference (human + AI, not part of the comparison) row comes from runs/reference. A human wrote it with AI help for A02 before this experiment, so it sits apart below the comparison and is never ranked. Its functional mark, hidden, unseen and acceptance grades, costs and audit are not measured here. The blind rating scored it as a calibration item in round-b, and its row shows those medians. Its tests live only in the hidden jar.
Reference
| Method | A02 reference (human + AI, not part of the comparison) |
|---|---|
| Branch | A02 origin/submission |
| Run order | not measured |
| Functional mark /15 | not measured |
| Hidden pass /76 | not measured |
| Unseen pass /69 | not measured |
| Acceptance pass /6 | not measured |
| Strict PMD compliance | pass |
| Mutation score | 154/165, 0 not run (the reference's own TestKalah tests from the hidden jar) |
| Static analysis violations | 0 |
| Lines of code | 380 (src only) |
| Tokens | not measured |
| Weighted tokens | not measured |
| APFDc-H | not measured |
| Cost to completion | not measured |
| M | not measured |
| M_io | not measured |
| Wall time | not measured |
| Classes | 10 |
| Methods | 55 |
| Fields | 19 |
| Bench median ms | 500 |
| Clear Presentation /5 | 4 |
| Explanatory Language /5 | 3 |
| Consistent Code /5 | 4 |
| Used Content /5 | 4 |
| Simple Constructs /5 | 4 |
| Minimal Duplication /5 | 4 |
| Modular Structure /5 | 4 |
| Problem Alignment (tentative) /5 | 4 |
| Agent calls | not measured |
| Audit flags | not measured |