ES-103 Evidence¶
Purpose¶
This page defines the evidence expected from ES-103.
Planning evidence demonstrates that the work required by ES-102 has been decomposed, sequenced, assigned, estimated, risk-managed, and reviewed. It also demonstrates that uncertainty has been made visible rather than hidden inside optimistic schedules.
Required evidence¶
| Artifact | Evidence Purpose |
|---|---|
planning_overview.md |
Connects requirements to planning assumptions and objectives. |
work_breakdown_structure.md |
Decomposes work into accountable, traceable work packages. |
milestones.md |
Defines evidence-backed readiness checkpoints. |
roles_and_responsibilities.md |
Assigns ownership and accountability. |
risk_register.md |
Identifies and manages planning and execution risks. |
dependency_map.md |
Shows dependency relationships and blockers. |
estimation_record.md |
Records effort estimates and uncertainty. |
planning_review.md |
Records review findings and corrections. |
planning_readiness_summary.md |
Summarizes readiness for ES-104. |
Evidence quality expectations¶
Planning evidence should be traceable to requirements and constraints, realistic, risk-aware, explicit about ownership, explicit about dependencies, honest about uncertainty, clear about evidence outputs, and useful to architecture work.
Planning evidence is weak when it only lists tasks. It is stronger when it shows why the work exists, what it produces, who owns it, what it depends on, and what would make it fail.
Work package quality test¶
A work package is strong enough when a reviewer can answer:
- What deliverable or outcome does this work produce?
- Which requirement, constraint, risk, review obligation, or evidence obligation justifies it?
- Who owns it?
- What does it depend on?
- What evidence will it produce?
- How large is the work, approximately?
- What could block it?
- What review or readiness point confirms it is complete?
Milestone quality test¶
A milestone is strong enough when it represents a meaningful readiness checkpoint.
A weak milestone says:
Architecture complete by Friday.
A stronger milestone says:
Architecture candidate reviewed against functional requirements, nonfunctional requirements, constraints, major risks, and dependency map; unresolved tradeoffs recorded.
The second milestone describes evidence and decision criteria.
Estimation quality test¶
An estimate is useful when it helps planning decisions without pretending to be precise.
Use ranges, effort bands, assumptions, and uncertainty notes. If the team cannot estimate responsibly, define an investigation task.
Risk evidence¶
Risk evidence should identify the risk, why it matters, likelihood, impact, owner, mitigation, trigger, and current status.
A risk without an owner is not managed. A risk without a trigger is easy to ignore. A risk without mitigation is only a concern, not a plan.
AI-use evidence¶
If AI materially assists planning, record how it was used and how the plan was reviewed.
Minimum recommended statement:
AI assistance was used to identify planning gaps, dependencies, risks, or sequencing concerns. The engineering team reviewed the final plan and remains responsible for estimates, ownership, sequencing, risk decisions, and commitments.
Review evidence¶
Planning should be reviewed before ES-104.
Review findings should be recorded in:
docs/planning/planning_review.md
The review should check missing work, unowned work, unrealistic estimates, hidden dependencies, missing evidence, unmanaged risks, and unsupported commitments.
Lifecycle chain¶
In ES-103, evidence supports a specific readiness decision:
Planning evidence
↓
Architecture readiness decision
↓
ES-104 begins with known work, risks, owners, dependencies, and constraints
A plan that does not include review and evidence work is incomplete. Engineering work includes producing proof that the work occurred responsibly.
Planning evidence should make later surprises less damaging, not pretend surprises will not happen.