| 1 | # DE/DX Playbook v1 |
| 2 | |
| 3 | ## Purpose |
| 4 | Practical standards for developer effectiveness and sustainable delivery speed. |
| 5 | |
| 6 | ## Core Outcomes |
| 7 | - Faster safe delivery with lower cognitive load. |
| 8 | - Predictable onboarding and fewer environment-specific failures. |
| 9 | - Higher signal in tooling, tests, and feedback loops. |
| 10 | |
| 11 | ## Evidence-Based Working Format |
| 12 | - Fact: document current observable state (metrics, incidents, queue age, failure class). |
| 13 | - Hypothesis: state expected causal change from an intervention. |
| 14 | - Check: define the smallest test and observation window. |
| 15 | - Decision criterion: predefine keep/rollback/escalate threshold. |
| 16 | - Confidence mark: explicitly tag confidence and uncertainty before rollout. |
| 17 | |
| 18 | ## Operating Principles |
| 19 | - Make the easy path the correct path. |
| 20 | - Minimize friction before adding process. |
| 21 | - Prefer automation over recurring manual rituals. |
| 22 | - Preserve focus: lower context switching and queue growth. |
| 23 | |
| 24 | ## Environment and Tooling Baseline |
| 25 | - One-command local start for the main product path. |
| 26 | - Explicit dependency and version policy. |
| 27 | - Stable debug profile with reproducible breakpoints. |
| 28 | - Project templates and scripts for repeated workflows. |
| 29 | |
| 30 | ## Delivery and Flow Rules |
| 31 | - Small batches, clear scope, reversible changes. |
| 32 | - WIP limits for active tasks. |
| 33 | - Fast feedback from lint/build/test before merge. |
| 34 | - Visible bottlenecks and queue size monitoring. |
| 35 | |
| 36 | ## Code Review Quality |
| 37 | - Review for risk and behavior, not style-only nits. |
| 38 | - Require rationale for architecture-impacting changes. |
| 39 | - Prefer actionable comments with concrete alternatives. |
| 40 | - Track recurring review findings into checklist updates. |
| 41 | |
| 42 | ## Reliability and Dev Loop |
| 43 | - Keep a minimal smoke suite for critical user paths. |
| 44 | - Define rollback path for risky releases. |
| 45 | - Capture incident lessons as reusable rules. |
| 46 | - Protect p95/p99 user-facing performance during iteration. |
| 47 | |
| 48 | ## Metrics |
| 49 | - Lead time for changes. |
| 50 | - Change failure rate. |
| 51 | - Mean time to restore. |
| 52 | - PR cycle time and review latency. |
| 53 | - Onboarding time to first successful contribution. |
| 54 | |
| 55 | ## Literature synthesis (IDE / UI — классика, не метрики команды) |
| 56 | |
| 57 | - **Опыт разработчика и интегрированная среда** (Osmani, Goldberg/Smalltalk-80, diSessa/Boxer): `kb-ide-dx-literature-evidence-v1.md` — принципы инструмента и IDE; **дополняет** этот playbook, но **не заменяет** разделы про поток поставки и DORA-ориентир выше. |
| 58 | - **UI/UX интерфейса** (Norman, Nielsen, Shneiderman, Krug): `kb-ui-ux-literature-evidence-v1.md` — эвристики и язык ревью; пересечение с доменом **HCI** (`status-hci-v1.md`). |
| 59 | |
| 60 | ## Revisit Triggers |
| 61 | - Rising cycle time or merge delay. |
| 62 | - Frequent environment-specific failures. |
| 63 | - Repeat incident classes with no checklist improvement. |
| 64 | - Team feedback that tooling is harder than product work. |
| 65 | |
| 66 | |