add-evaluation-tool

Strategic Plan

Project goals

• Deliver an end-to-end architecture evaluation tool that unifies Attribute-Driven Design guidance, matrix-based analysis, and visual diffing of PlantUML models.
• Provide a repeatable workflow for student and research teams to capture architectural decisions, evaluate non-functional requirements, and compare architecture versions over time.
• Build an open, extensible foundation (frontend, backend, data model) that can evolve into a production-ready platform after the course.

Threshold of success

• MVP users can upload a PlantUML component diagram and automatically receive a populated NFR × Component evaluation matrix within 2 minutes.
• The tool stores at least five historical versions of an architecture and produces a diff report (diagram + scores) in under 5 seconds per comparison.
• At least three external testers (students or researchers) can complete the core workflow—upload diagram, adjust scores, compare versions—without direct assistance.
• Documentation covers onboarding (README), architecture, and usage scenarios, enabling new contributors to add features within one sprint.

Feature roadmap

• Sprint 0: team formation, research, interview, documented findings, strategic plan.
• Sprint 1: refine product backlog, finalize architecture decisions, and stub core services.
• Sprint 2: implement baseline PlantUML ingestion and persistence layer with seed data.
• Sprint 3: deliver first-cut matrix UI and hook it to parsed model data.
• Sprint 4: add scoring logic, version storage APIs, and happy-path evaluation flow.
• Sprint 5: prototype diff visualizations (textual + simple diagram highlights) and gather feedback.
• Sprint 6: implement user authorization.
• Sprint 7: expand reporting, accessibility, and onboarding documentation based on user feedback.
• Sprint 8: finalize release readiness—security review, deployment guides, and course presentation prep.

Progress monitoring

• Sprint reviews every week with demo-ready increments and acceptance criteria checklists.
• KPI dashboard (GitHub project board, burndown chart) monitoring story throughput, open issues, and test coverage.
• Automated CI pipeline running unit/lint tests on every PR; nightly builds executing integration suites and performance smoke tests.
• Every week stakeholder sync with customer to validate assumptions, confirm priorities, and collect qualitative feedback.

Contingency plans

• If PlantUML parsing proves unreliable, pivot to Structurizr DSL import with automatic conversion to PlantUML, keeping the matrix workflow intact.
• Should scoring formulas remain undefined, ship with a configurable default (weighted averages) and expose admin settings to adjust once final guidance arrives.
• For visual diff complexity, release a phased approach: start with textual diff + matrix delta, add diagram overlays once core value is stable.
• If scope threatens schedule, lock MVP at matrix evaluation + version comparison and defer advanced automation (ADD guidance, AI hints) to future iterations.

This site is open source. Improve this page.