Assessment¶
The following is a suggested research-oriented model, not a claim about an official university grading scheme.
| Component | Suggested weight | Main evidence |
|---|---|---|
| A1 — Principle-Driven Design Review | 10% | analytical reasoning |
| A2 — Pattern & Anti-pattern Case Study | 10% | contextual pattern judgment |
| A3 — Refactoring to Patterns | 15% | transformation quality and safety |
| A4 — Automated Design Analysis | 15% | implementation + evaluation |
| Seminar / paper leadership | 10% | research communication |
| Research project | 35% | novelty/replication, rigor, artifact, report |
| Participation / design clinics | 5% | informed technical discussion |
What is rewarded¶
Across assignments, the grading emphasis should be on:
- explicit assumptions;
- evidence-backed design reasoning;
- correct use of vocabulary;
- quality of alternatives considered;
- behavior preservation and tests;
- reproducibility of automated experiments;
- critical analysis of limitations.
What is not rewarded by itself¶
- using many patterns;
- increasing the number of classes/interfaces;
- lowering one metric while damaging the design elsewhere;
- reporting only aggregate accuracy without error analysis;
- applying an LLM and calling the output “automated design” without a baseline or evaluation protocol.