Martin Fowler’s website

• → O031 · Context Engineering: Bringing Engineering Discipline to Prompts
  F_EDITORIAL_COMPARISON · edge f00359 · key MF_V12:F_EDITORIAL_COMPARISON

These retain the graph’s exact relation labels. A connection is not automatically a causal claim or an independent replication.

{
  "label": "The toolchain",
  "kind": "practitioner_essay",
  "collection": "F",
  "authors": "Birgitta Böckeler",
  "authors_json": "[\"Birgitta Böckeler\"]",
  "publication_date": "2023-07-26",
  "date_provenance": "Dated entry on the supplied series index; article date also used where printed",
  "url": "https://martinfowler.com/articles/exploring-gen-ai/01-tool-landscape.html",
  "publisher": "martinfowler.com",
  "series": "Exploring Generative AI",
  "series_url": "https://martinfowler.com/articles/exploring-gen-ai.html",
  "snapshot_date": "2026-10-13",
  "contribution": "Organizes coding-assistance tools by task, interaction style, prompt construction and model or hosting choices. The framework helps separate what a tool does from how a developer supplies its context.",
  "limitation": "A July 2023 conceptual landscape, not a comparative benchmark or a current product inventory.",
  "evidence_class": "Conceptual tool taxonomy",
  "source_locator": "Body: task and interaction categories; prompt composition; models and hosting",
  "verification_scope": "Primary article byline and selected body passages inspected; no independent execution or statistical replication",
  "original_data_replicated": false,
  "new_manuscript_citation": false,
  "source_body_archived": false,
  "primary_stream": "T:Engineering structure",
  "streams_json": "[\"T:Engineering structure\", \"T:Productivity\"]",
  "topics_json": "[\"K:Developer workflow and adoption\", \"K:Agent context and persistent memory\"]",
  "editorial_chapters_json": "[\"C:C01\", \"C:C03\", \"C:C09\"]",
  "mapping_status": "Independent editorial mapping, not classification by the article authors or SE@PoliMi",
  "book_citation_status": "COMPANION_READING_NOT_IN_BOOK",
  "node_class": "FOWLER_SERIES_ADDITIVE_OVERLAY_V12",
  "ndse2_actual_chapters_json": "[]",
  "ndse2_in_bibliography": false,
  "ndse2_citation_basis": "Final 13.7.1 active EN/ITA source tree; index links and nocite are not chapter citations",
  "historical_v12_book_citation_status": "NOT_ADDED_TO_BOOK"
}

• ← F003 · In-line assistance - when is it more useful?
  F_BODY_LINKS_TO · edge f00030 · key MF_V12:F_BODY_LINKS_TO
• ← F008 · How to tackle unreliability of coding assistants
  F_BODY_LINKS_TO · edge f00081 · key MF_V12:F_BODY_LINKS_TO
• → W077 · Evaluating and Mitigating the Misguidance Effect of Buggy Code in LLM-Generated Unit Tests
  F_EDITORIAL_COMPARISON · edge f00360 · key MF_V12:F_EDITORIAL_COMPARISON

These retain the graph’s exact relation labels. A connection is not automatically a causal claim or an independent replication.

{
  "label": "Median - A tale in three functions",
  "kind": "practitioner_essay",
  "collection": "F",
  "authors": "Birgitta Böckeler",
  "authors_json": "[\"Birgitta Böckeler\"]",
  "publication_date": "2023-07-27",
  "date_provenance": "Dated entry on the supplied series index; article date also used where printed",
  "url": "https://martinfowler.com/articles/exploring-gen-ai/02-median-tale.html",
  "publisher": "martinfowler.com",
  "series": "Exploring Generative AI",
  "series_url": "https://martinfowler.com/articles/exploring-gen-ai.html",
  "snapshot_date": "2026-10-13",
  "contribution": "Examines three generated median implementations and the defects or side effects that a knowledgeable developer must recognize. The example makes the choice of test expectations part of the engineering work.",
  "limitation": "One small TypeScript example; neither defect prevalence nor a general model-quality comparison was established.",
  "evidence_class": "Worked programming example",
  "source_locator": "Body: the three median implementations and the test discussion",
  "verification_scope": "Primary article byline and selected body passages inspected; no independent execution or statistical replication",
  "original_data_replicated": false,
  "new_manuscript_citation": false,
  "source_body_archived": false,
  "primary_stream": "T:Verification & review",
  "streams_json": "[\"T:Verification & review\", \"T:Code quality & security\"]",
  "topics_json": "[\"K:Independent verification and oracles\", \"K:Automation bias and trust\"]",
  "editorial_chapters_json": "[\"C:C05\", \"C:C07\", \"C:C10\"]",
  "mapping_status": "Independent editorial mapping, not classification by the article authors or SE@PoliMi",
  "book_citation_status": "COMPANION_READING_NOT_IN_BOOK",
  "node_class": "FOWLER_SERIES_ADDITIVE_OVERLAY_V12",
  "ndse2_actual_chapters_json": "[]",
  "ndse2_in_bibliography": false,
  "ndse2_citation_basis": "Final 13.7.1 active EN/ITA source tree; index links and nocite are not chapter citations",
  "historical_v12_book_citation_status": "NOT_ADDED_TO_BOOK"
}

• → F002 · Median - A tale in three functions
  F_BODY_LINKS_TO · edge f00030 · key MF_V12:F_BODY_LINKS_TO
• → W010 · Writing Code vs. Shipping Code: Productivity Effects Across Generations of AI Coding Tools
  F_EDITORIAL_COMPARISON · edge f00361 · key MF_V12:F_EDITORIAL_COMPARISON

These retain the graph’s exact relation labels. A connection is not automatically a causal claim or an independent replication.

{
  "label": "In-line assistance - when is it more useful?",
  "kind": "practitioner_essay",
  "collection": "F",
  "authors": "Birgitta Böckeler",
  "authors_json": "[\"Birgitta Böckeler\"]",
  "publication_date": "2023-08-01",
  "date_provenance": "Dated entry on the supplied series index; article date also used where printed",
  "url": "https://martinfowler.com/articles/exploring-gen-ai/03-coding-assistance-when-useful.html",
  "publisher": "martinfowler.com",
  "series": "Exploring Generative AI",
  "series_url": "https://martinfowler.com/articles/exploring-gen-ai.html",
  "snapshot_date": "2026-10-13",
  "contribution": "Discusses when inline suggestions help, considering task familiarity, technology, change size, developer experience and the effort of checking or reworking the output.",
  "limitation": "Experience-based guidance, largely from experienced developers; usefulness is not a measured universal productivity effect.",
  "evidence_class": "Practitioner experience",
  "source_locator": "Body: usefulness factors and review/rework effort",
  "verification_scope": "Primary article byline and selected body passages inspected; no independent execution or statistical replication",
  "original_data_replicated": false,
  "new_manuscript_citation": false,
  "source_body_archived": false,
  "primary_stream": "T:Productivity",
  "streams_json": "[\"T:Productivity\", \"T:Verification & review\"]",
  "topics_json": "[\"K:Developer workflow and adoption\", \"K:Human cognitive load\", \"K:Economic costs and return\"]",
  "editorial_chapters_json": "[\"C:C02\", \"C:C03\", \"C:C10\", \"C:C12\"]",
  "mapping_status": "Independent editorial mapping, not classification by the article authors or SE@PoliMi",
  "book_citation_status": "COMPANION_READING_NOT_IN_BOOK",
  "node_class": "FOWLER_SERIES_ADDITIVE_OVERLAY_V12",
  "ndse2_actual_chapters_json": "[]",
  "ndse2_in_bibliography": false,
  "ndse2_citation_basis": "Final 13.7.1 active EN/ITA source tree; index links and nocite are not chapter citations",
  "historical_v12_book_citation_status": "NOT_ADDED_TO_BOOK"
}

• ← F005 · Coding assistants do not replace pair programming
  F_BODY_LINKS_TO · edge f00052 · key MF_V12:F_BODY_LINKS_TO
• → O019 · Code Review in the Age of AI
  F_EDITORIAL_COMPARISON · edge f00362 · key MF_V12:F_EDITORIAL_COMPARISON

These retain the graph’s exact relation labels. A connection is not automatically a causal claim or an independent replication.

{
  "label": "In-line assistance - how can it get in the way?",
  "kind": "practitioner_essay",
  "collection": "F",
  "authors": "Birgitta Böckeler",
  "authors_json": "[\"Birgitta Böckeler\"]",
  "publication_date": "2023-08-03",
  "date_provenance": "Dated entry on the supplied series index; article date also used where printed",
  "url": "https://martinfowler.com/articles/exploring-gen-ai/04-coding-assistance-how-in-the-way.html",
  "publisher": "martinfowler.com",
  "series": "Exploring Generative AI",
  "series_url": "https://martinfowler.com/articles/exploring-gen-ai.html",
  "snapshot_date": "2026-10-13",
  "contribution": "Describes assistants repeating poor patterns from their context and the burden of continually judging suggestions. Relates these experiences to automation bias, anchoring and reluctance to discard generated work.",
  "limitation": "Qualitative examples and interpretive mechanisms, not measured frequencies of bias or review fatigue.",
  "evidence_class": "Practitioner experience",
  "source_locator": "Body: reinforcing existing practices; reviewing instead of producing",
  "verification_scope": "Primary article byline and selected body passages inspected; no independent execution or statistical replication",
  "original_data_replicated": false,
  "new_manuscript_citation": false,
  "source_body_archived": false,
  "primary_stream": "T:Verification & review",
  "streams_json": "[\"T:Verification & review\", \"T:System knowledge & memory\", \"T:Code quality & security\"]",
  "topics_json": "[\"K:Automation bias and trust\", \"K:Human cognitive load\", \"K:Agent context and persistent memory\"]",
  "editorial_chapters_json": "[\"C:C05\", \"C:C08\", \"C:C09\"]",
  "mapping_status": "Independent editorial mapping, not classification by the article authors or SE@PoliMi",
  "book_citation_status": "COMPANION_READING_NOT_IN_BOOK",
  "node_class": "FOWLER_SERIES_ADDITIVE_OVERLAY_V12",
  "ndse2_actual_chapters_json": "[]",
  "ndse2_in_bibliography": false,
  "ndse2_citation_basis": "Final 13.7.1 active EN/ITA source tree; index links and nocite are not chapter citations",
  "historical_v12_book_citation_status": "NOT_ADDED_TO_BOOK"
}

• → F004 · In-line assistance - how can it get in the way?
  F_BODY_LINKS_TO · edge f00052 · key MF_V12:F_BODY_LINKS_TO
• → W173 · Quantifying and Mitigating Turnover-Induced Knowledge Loss: Case Studies of Chrome and a Project at Avaya
  F_EDITORIAL_COMPARISON · edge f00363 · key MF_V12:F_EDITORIAL_COMPARISON

These retain the graph’s exact relation labels. A connection is not automatically a causal claim or an independent replication.

{
  "label": "Coding assistants do not replace pair programming",
  "kind": "practitioner_essay",
  "collection": "F",
  "authors": "Birgitta Böckeler",
  "authors_json": "[\"Birgitta Böckeler\"]",
  "publication_date": "2023-08-10",
  "date_provenance": "Dated entry on the supplied series index; article date also used where printed",
  "url": "https://martinfowler.com/articles/exploring-gen-ai/05-not-your-pair-programmer.html",
  "publisher": "martinfowler.com",
  "series": "Exploring Generative AI",
  "series_url": "https://martinfowler.com/articles/exploring-gen-ai.html",
  "snapshot_date": "2026-10-13",
  "contribution": "Argues that code assistance does not replace the human knowledge sharing, collaboration and collective understanding associated with pair programming. Tools can instead be used by a human pair.",
  "limitation": "An argument about the functions of pairing; not a controlled comparison of pairing arrangements.",
  "evidence_class": "Conceptual argument",
  "source_locator": "Body: knowledge sharing, team flow and pairing with assistance",
  "verification_scope": "Primary article byline and selected body passages inspected; no independent execution or statistical replication",
  "original_data_replicated": false,
  "new_manuscript_citation": false,
  "source_body_archived": false,
  "primary_stream": "T:Skills & learning",
  "streams_json": "[\"T:Skills & learning\", \"T:Engineering structure\", \"T:System knowledge & memory\"]",
  "topics_json": "[\"K:Human-agent collaboration\", \"K:Ownership and knowledge transfer\", \"K:Skills formation and teaching\"]",
  "editorial_chapters_json": "[\"C:C10\", \"C:C11\"]",
  "mapping_status": "Independent editorial mapping, not classification by the article authors or SE@PoliMi",
  "book_citation_status": "COMPANION_READING_NOT_IN_BOOK",
  "node_class": "FOWLER_SERIES_ADDITIVE_OVERLAY_V12",
  "ndse2_actual_chapters_json": "[]",
  "ndse2_in_bibliography": false,
  "ndse2_citation_basis": "Final 13.7.1 active EN/ITA source tree; index links and nocite are not chapter citations",
  "historical_v12_book_citation_status": "NOT_ADDED_TO_BOOK"
}

• → B014 · PART II: Quality Assurance
  F_EDITORIAL_COMPARISON · edge f00364 · key MF_V12:F_EDITORIAL_COMPARISON
• ← F031 · TDD inside the agent loop - theater or actual value?
  F_EDITORIAL_COMPARISON · edge f00398 · key MF_V12:F_EDITORIAL_COMPARISON

These retain the graph’s exact relation labels. A connection is not automatically a causal claim or an independent replication.

{
  "label": "TDD with GitHub Copilot",
  "kind": "practitioner_essay",
  "collection": "F",
  "authors": "Paul Sobocinski",
  "authors_json": "[\"Paul Sobocinski\"]",
  "publication_date": "2023-08-17",
  "date_provenance": "Dated entry on the supplied series index; article date also used where printed",
  "url": "https://martinfowler.com/articles/exploring-gen-ai/06-tdd-with-coding-assistance.html",
  "publisher": "martinfowler.com",
  "series": "Exploring Generative AI",
  "series_url": "https://martinfowler.com/articles/exploring-gen-ai.html",
  "snapshot_date": "2026-10-13",
  "contribution": "Explores assistance during red, green and refactor steps. Human test intent and assertions remain important; suggestions can help implementation while also overproducing code or struggling with refactoring.",
  "limitation": "Experience with a human-led workflow and an early tool generation; not evidence about fully autonomous TDD or a causal quality gain.",
  "evidence_class": "Practitioner workflow report",
  "source_locator": "Body: red, green and refactoring discussions",
  "verification_scope": "Primary article byline and selected body passages inspected; no independent execution or statistical replication",
  "original_data_replicated": false,
  "new_manuscript_citation": false,
  "source_body_archived": false,
  "primary_stream": "T:Verification & review",
  "streams_json": "[\"T:Verification & review\", \"T:Skills & learning\"]",
  "topics_json": "[\"K:Independent verification and oracles\", \"K:Human-agent collaboration\"]",
  "editorial_chapters_json": "[\"C:C03\", \"C:C07\", \"C:C10\"]",
  "mapping_status": "Independent editorial mapping, not classification by the article authors or SE@PoliMi",
  "book_citation_status": "COMPANION_READING_NOT_IN_BOOK",
  "node_class": "FOWLER_SERIES_ADDITIVE_OVERLAY_V12",
  "ndse2_actual_chapters_json": "[]",
  "ndse2_in_bibliography": false,
  "ndse2_citation_basis": "Final 13.7.1 active EN/ITA source tree; index links and nocite are not chapter citations",
  "historical_v12_book_citation_status": "NOT_ADDED_TO_BOOK"
}

• ← F019 · I still care about the code
  F_BODY_LINKS_TO · edge f00197 · key MF_V12:F_BODY_LINKS_TO
• → B050 · The Process: Four Phases, Four Gates
  F_EDITORIAL_COMPARISON · edge f00365 · key MF_V12:F_EDITORIAL_COMPARISON

These retain the graph’s exact relation labels. A connection is not automatically a causal claim or an independent replication.

{
  "label": "How is GenAI different from other code generators?",
  "kind": "practitioner_essay",
  "collection": "F",
  "authors": "Birgitta Böckeler",
  "authors_json": "[\"Birgitta Böckeler\"]",
  "publication_date": "2023-09-19",
  "date_provenance": "Dated entry on the supplied series index; article date also used where printed",
  "url": "https://martinfowler.com/articles/exploring-gen-ai/07-how-is-this-different.html",
  "publisher": "martinfowler.com",
  "series": "Exploring Generative AI",
  "series_url": "https://martinfowler.com/articles/exploring-gen-ai.html",
  "snapshot_date": "2026-10-13",
  "contribution": "Contrasts generative assistance with model-driven development and domain-specific code generators. Examines how control and abstraction differ when generation draws on broad learned patterns rather than an explicitly designed transformation.",
  "limitation": "Conceptual comparison; it does not demonstrate the reliability or economic superiority of either approach.",
  "evidence_class": "Conceptual argument",
  "source_locator": "Body: abstraction levels and comparison with traditional generation",
  "verification_scope": "Primary article byline and selected body passages inspected; no independent execution or statistical replication",
  "original_data_replicated": false,
  "new_manuscript_citation": false,
  "source_body_archived": false,
  "primary_stream": "T:Engineering structure",
  "streams_json": "[\"T:Engineering structure\"]",
  "topics_json": "[\"K:Specifications and governance\", \"K:Architecture and modularity\"]",
  "editorial_chapters_json": "[\"C:C01\", \"C:C04\", \"C:C05\"]",
  "mapping_status": "Independent editorial mapping, not classification by the article authors or SE@PoliMi",
  "book_citation_status": "COMPANION_READING_NOT_IN_BOOK",
  "node_class": "FOWLER_SERIES_ADDITIVE_OVERLAY_V12",
  "ndse2_actual_chapters_json": "[]",
  "ndse2_in_bibliography": false,
  "ndse2_citation_basis": "Final 13.7.1 active EN/ITA source tree; index links and nocite are not chapter citations",
  "historical_v12_book_citation_status": "NOT_ADDED_TO_BOOK"
}

• → F002 · Median - A tale in three functions
  F_BODY_LINKS_TO · edge f00081 · key MF_V12:F_BODY_LINKS_TO
• → B109 · Choosing the Right Loop
  F_EDITORIAL_COMPARISON · edge f00366 · key MF_V12:F_EDITORIAL_COMPARISON
• ← NDSE2:C03 · Limits and Capability of Coding Agents
  NDSE2_CHAPTER_CITES · edge u20261018_00208 · key UPDATE_20261018:NDSE2_CHAPTER_CITES:EN
• ← NDSE2:C03 · Limits and Capability of Coding Agents
  NDSE2_CHAPTER_CITES · edge u20261018_00229 · key UPDATE_20261018:NDSE2_CHAPTER_CITES:ITA

These retain the graph’s exact relation labels. A connection is not automatically a causal claim or an independent replication.

{
  "label": "How to tackle unreliability of coding assistants",
  "kind": "practitioner_essay",
  "collection": "F",
  "authors": "Birgitta Böckeler",
  "authors_json": "[\"Birgitta Böckeler\"]",
  "publication_date": "2023-11-29",
  "date_provenance": "Dated entry on the supplied series index; article date also used where printed",
  "url": "https://martinfowler.com/articles/exploring-gen-ai/08-how-to-tackle-unreliability.html",
  "publisher": "martinfowler.com",
  "series": "Exploring Generative AI",
  "series_url": "https://martinfowler.com/articles/exploring-gen-ai.html",
  "snapshot_date": "2026-10-13",
  "contribution": "Frames confidence in assistance around feedback, the developer’s ability to check an answer and the consequences of error. Proposes timeboxing unproductive assistance and retaining responsibility for the result.",
  "limitation": "Personal heuristics rather than calibrated reliability estimates; informal numerical intuitions are not treated as measured rates.",
  "evidence_class": "Practitioner guidance",
  "source_locator": "Body: confidence, timeboxing and assistant persona",
  "verification_scope": "Primary article byline and selected body passages inspected; no independent execution or statistical replication",
  "original_data_replicated": false,
  "new_manuscript_citation": true,
  "source_body_archived": false,
  "primary_stream": "T:Verification & review",
  "streams_json": "[\"T:Verification & review\", \"T:What agents can do\"]",
  "topics_json": "[\"K:Agent failure and escalation\", \"K:Independent verification and oracles\", \"K:Automation bias and trust\"]",
  "editorial_chapters_json": "[\"C:C03\", \"C:C07\", \"C:C08\", \"C:C13\"]",
  "mapping_status": "Independent editorial mapping, not classification by the article authors or SE@PoliMi",
  "book_citation_status": "IN_PUBLISHED_NDSE2_BIBLIOGRAPHY",
  "node_class": "FOWLER_SERIES_ADDITIVE_OVERLAY_V12",
  "ndse2_actual_chapters_json": "[\"C:C03\"]",
  "ndse2_in_bibliography": true,
  "ndse2_citation_basis": "Final 13.7.1 active EN/ITA source tree; index links and nocite are not chapter citations",
  "historical_v12_book_citation_status": "NOT_ADDED_TO_BOOK",
  "historical_v12_new_manuscript_citation": false
}

• → W155 · ChatGPT: Friend or Foe When Comprehending and Changing Unfamiliar Code
  F_EDITORIAL_COMPARISON · edge f00367 · key MF_V12:F_EDITORIAL_COMPARISON
• → O002 · Brownfield Agentic Engineering
  F_EDITORIAL_COMPARISON · edge f00368 · key MF_V12:F_EDITORIAL_COMPARISON
• ← F035 · When Software Meets Hardware
  F_EDITORIAL_COMPARISON · edge u20261018_00031 · key UPDATE_20261018:F_EDITORIAL_COMPARISON

These retain the graph’s exact relation labels. A connection is not automatically a causal claim or an independent replication.

{
  "label": "Onboarding to a 'legacy' codebase with the help of AI",
  "kind": "practitioner_essay",
  "collection": "F",
  "authors": "Birgitta Böckeler",
  "authors_json": "[\"Birgitta Böckeler\"]",
  "publication_date": "2024-08-15",
  "date_provenance": "Dated entry on the supplied series index; article date also used where printed",
  "url": "https://martinfowler.com/articles/exploring-gen-ai/09-ai-help-onboarding-codebase.html",
  "publisher": "martinfowler.com",
  "series": "Exploring Generative AI",
  "series_url": "https://martinfowler.com/articles/exploring-gen-ai.html",
  "snapshot_date": "2026-10-13",
  "contribution": "Recounts work on an unfamiliar Bahmni/OpenMRS issue. AI helped with terminology and code navigation, but environment setup, fixtures and mocking remained obstacles, and the issue was not solved in the attempt.",
  "limitation": "One exploratory onboarding attempt; no measured reduction in onboarding time or demonstrated long-term comprehension.",
  "evidence_class": "Exploratory case report",
  "source_locator": "Body: issue exploration, code search, environment and testing attempts",
  "verification_scope": "Primary article byline and selected body passages inspected; no independent execution or statistical replication",
  "original_data_replicated": false,
  "new_manuscript_citation": false,
  "source_body_archived": false,
  "primary_stream": "T:System knowledge & memory",
  "streams_json": "[\"T:System knowledge & memory\", \"T:Skills & learning\"]",
  "topics_json": "[\"K:Onboarding and developer fluency\", \"K:System understanding and rationale\"]",
  "editorial_chapters_json": "[\"C:C05\", \"C:C09\", \"C:C10\"]",
  "mapping_status": "Independent editorial mapping, not classification by the article authors or SE@PoliMi",
  "book_citation_status": "COMPANION_READING_NOT_IN_BOOK",
  "node_class": "FOWLER_SERIES_ADDITIVE_OVERLAY_V12",
  "ndse2_actual_chapters_json": "[]",
  "ndse2_in_bibliography": false,
  "ndse2_citation_basis": "Final 13.7.1 active EN/ITA source tree; index links and nocite are not chapter citations",
  "historical_v12_book_citation_status": "NOT_ADDED_TO_BOOK"
}

• → W101 · Articulate but Wrong: Self-Review Failures in LLM-Based Code Modernization
  F_EDITORIAL_COMPARISON · edge f00369 · key MF_V12:F_EDITORIAL_COMPARISON

These retain the graph’s exact relation labels. A connection is not automatically a causal claim or an independent replication.

{
  "label": "Building an AI agent application to migrate a tech stack",
  "kind": "practitioner_essay",
  "collection": "F",
  "authors": "Birgitta Böckeler",
  "authors_json": "[\"Birgitta Böckeler\"]",
  "publication_date": "2024-08-20",
  "date_provenance": "Dated entry on the supplied series index; article date also used where printed",
  "url": "https://martinfowler.com/articles/exploring-gen-ai/10-ai-for-tech-stack-migration.html",
  "publisher": "martinfowler.com",
  "series": "Exploring Generative AI",
  "series_url": "https://martinfowler.com/articles/exploring-gen-ai.html",
  "snapshot_date": "2026-10-13",
  "contribution": "Builds a tool-using agent to migrate tests from Enzyme to React Testing Library. The account exposes file-editing, test-feedback and context-management problems alongside successful transformations.",
  "limitation": "A bounded migration experiment with substantial failures; not proof of reliable autonomous migration across a codebase.",
  "evidence_class": "Exploratory implementation experiment",
  "source_locator": "Body: manually solved baseline, agent setup and migration observations",
  "verification_scope": "Primary article byline and selected body passages inspected; no independent execution or statistical replication",
  "original_data_replicated": false,
  "new_manuscript_citation": false,
  "source_body_archived": false,
  "primary_stream": "T:What agents can do",
  "streams_json": "[\"T:What agents can do\", \"T:Code quality & security\", \"T:System knowledge & memory\"]",
  "topics_json": "[\"K:Agent failure and escalation\", \"K:Long-running agent execution\", \"K:Independent verification and oracles\"]",
  "editorial_chapters_json": "[\"C:C03\", \"C:C05\", \"C:C07\", \"C:C09\"]",
  "mapping_status": "Independent editorial mapping, not classification by the article authors or SE@PoliMi",
  "book_citation_status": "COMPANION_READING_NOT_IN_BOOK",
  "node_class": "FOWLER_SERIES_ADDITIVE_OVERLAY_V12",
  "ndse2_actual_chapters_json": "[]",
  "ndse2_in_bibliography": false,
  "ndse2_citation_basis": "Final 13.7.1 active EN/ITA source tree; index links and nocite are not chapter citations",
  "historical_v12_book_citation_status": "NOT_ADDED_TO_BOOK"
}

• ← F029 · Experiences with local models for coding
  F_BODY_LINKS_TO · edge f00310 · key MF_V12:F_BODY_LINKS_TO
• → W073 · The Effect of Complexity and Provenance on Code Review Decisions: Evidence from a Controlled Experiment
  F_EDITORIAL_COMPARISON · edge f00370 · key MF_V12:F_EDITORIAL_COMPARISON

These retain the graph’s exact relation labels. A connection is not automatically a causal claim or an independent replication.

{
  "label": "Expanding the solution size with multi-file editing",
  "kind": "practitioner_essay",
  "collection": "F",
  "authors": "Birgitta Böckeler",
  "authors_json": "[\"Birgitta Böckeler\"]",
  "publication_date": "2024-11-19",
  "date_provenance": "Dated entry on the supplied series index; article date also used where printed",
  "url": "https://martinfowler.com/articles/exploring-gen-ai/11-multi-file-editing.html",
  "publisher": "martinfowler.com",
  "series": "Exploring Generative AI",
  "series_url": "https://martinfowler.com/articles/exploring-gen-ai.html",
  "snapshot_date": "2026-10-13",
  "contribution": "Examines how multi-file edits enlarge the unit of AI-assisted work. Larger suggestions create new checking demands, including unrelated changes, and motivate layered feedback and comprehensible module boundaries.",
  "limitation": "A tool-use snapshot and qualitative experience; no comparative defect rate or current feature guarantee.",
  "evidence_class": "Practitioner tool exploration",
  "source_locator": "Body: multi-file mode, feedback loops and implications of larger changes",
  "verification_scope": "Primary article byline and selected body passages inspected; no independent execution or statistical replication",
  "original_data_replicated": false,
  "new_manuscript_citation": false,
  "source_body_archived": false,
  "primary_stream": "T:Code quality & security",
  "streams_json": "[\"T:Code quality & security\", \"T:Verification & review\", \"T:What agents can do\"]",
  "topics_json": "[\"K:Architecture and modularity\", \"K:Review capacity and maintainers\", \"K:Developer workflow and adoption\"]",
  "editorial_chapters_json": "[\"C:C03\", \"C:C05\", \"C:C08\"]",
  "mapping_status": "Independent editorial mapping, not classification by the article authors or SE@PoliMi",
  "book_citation_status": "COMPANION_READING_NOT_IN_BOOK",
  "node_class": "FOWLER_SERIES_ADDITIVE_OVERLAY_V12",
  "ndse2_actual_chapters_json": "[]",
  "ndse2_in_bibliography": false,
  "ndse2_citation_basis": "Final 13.7.1 active EN/ITA source tree; index links and nocite are not chapter citations",
  "historical_v12_book_citation_status": "NOT_ADDED_TO_BOOK"
}

• → O006 · Practical Loop Engineering
  F_EDITORIAL_COMPARISON · edge f00371 · key MF_V12:F_EDITORIAL_COMPARISON

These retain the graph’s exact relation labels. A connection is not automatically a causal claim or an independent replication.

{
  "label": "What role does LLM reasoning play for software tasks?",
  "kind": "practitioner_essay",
  "collection": "F",
  "authors": "Birgitta Böckeler",
  "authors_json": "[\"Birgitta Böckeler\"]",
  "publication_date": "2025-02-18",
  "date_provenance": "Dated entry on the supplied series index; article date also used where printed",
  "url": "https://martinfowler.com/articles/exploring-gen-ai/12-role-of-reasoning.html",
  "publisher": "martinfowler.com",
  "series": "Exploring Generative AI",
  "series_url": "https://martinfowler.com/articles/exploring-gen-ai.html",
  "snapshot_date": "2026-10-13",
  "contribution": "Questions where model reasoning adds value to coding tasks compared with supplying repository context and useful tool feedback. Considers debugging and planning as possible settings for investigation.",
  "limitation": "Tentative analysis from February 2025, not a current model-feature comparison or evidence that reasoning always helps or harms coding.",
  "evidence_class": "Conceptual analysis",
  "source_locator": "Body: reasoning models, software tasks and contextual/tool feedback",
  "verification_scope": "Primary article byline and selected body passages inspected; no independent execution or statistical replication",
  "original_data_replicated": false,
  "new_manuscript_citation": false,
  "source_body_archived": false,
  "primary_stream": "T:What agents can do",
  "streams_json": "[\"T:What agents can do\", \"T:System knowledge & memory\"]",
  "topics_json": "[\"K:Evaluation methodology and baselines\", \"K:Agent context and persistent memory\"]",
  "editorial_chapters_json": "[\"C:C03\", \"C:C09\"]",
  "mapping_status": "Independent editorial mapping, not classification by the article authors or SE@PoliMi",
  "book_citation_status": "COMPANION_READING_NOT_IN_BOOK",
  "node_class": "FOWLER_SERIES_ADDITIVE_OVERLAY_V12",
  "ndse2_actual_chapters_json": "[]",
  "ndse2_in_bibliography": false,
  "ndse2_citation_basis": "Final 13.7.1 active EN/ITA source tree; index links and nocite are not chapter citations",
  "historical_v12_book_citation_status": "NOT_ADDED_TO_BOOK"
}

• → P047 · Learnings from two years of using AI tools for software engineering
  F_EDITORIAL_COMPARISON · edge f00372 · key MF_V12:F_EDITORIAL_COMPARISON
• ← NDSE2:C10 · Developer Learning, Mentoring and Comprehension Debt
  NDSE2_CHAPTER_CITES · edge u20261018_00880 · key UPDATE_20261018:NDSE2_CHAPTER_CITES:EN
• ← NDSE2:C10 · Developer Learning, Mentoring and Comprehension Debt
  NDSE2_CHAPTER_CITES · edge u20261018_00903 · key UPDATE_20261018:NDSE2_CHAPTER_CITES:ITA

These retain the graph’s exact relation labels. A connection is not automatically a causal claim or an independent replication.

{
  "label": "The role of developer skills in agentic coding",
  "kind": "practitioner_essay",
  "collection": "F",
  "authors": "Birgitta Böckeler",
  "authors_json": "[\"Birgitta Böckeler\"]",
  "publication_date": "2025-03-25",
  "date_provenance": "Dated entry on the supplied series index; article date also used where printed",
  "url": "https://martinfowler.com/articles/exploring-gen-ai/13-role-of-developer-skills.html",
  "publisher": "martinfowler.com",
  "series": "Exploring Generative AI",
  "series_url": "https://martinfowler.com/articles/exploring-gen-ai.html",
  "snapshot_date": "2026-10-13",
  "contribution": "Uses agent missteps to explain why decomposition, debugging, design judgment and checking remain valuable. Separates immediate completion from team iteration and the lifetime of the codebase.",
  "limitation": "Experience-based examples, not a measured skills intervention or a forecast of developer employment.",
  "evidence_class": "Practitioner analysis",
  "source_locator": "Body: missteps and the time-to-commit, iteration and codebase-lifetime perspectives",
  "verification_scope": "Primary article byline and selected body passages inspected; no independent execution or statistical replication",
  "original_data_replicated": false,
  "new_manuscript_citation": true,
  "source_body_archived": false,
  "primary_stream": "T:Skills & learning",
  "streams_json": "[\"T:Skills & learning\", \"T:Code quality & security\", \"T:Engineering structure\"]",
  "topics_json": "[\"K:Skills formation and teaching\", \"K:Technical debt and complexity\", \"K:Human-agent collaboration\"]",
  "editorial_chapters_json": "[\"C:C01\", \"C:C05\", \"C:C10\", \"C:C11\"]",
  "mapping_status": "Independent editorial mapping, not classification by the article authors or SE@PoliMi",
  "book_citation_status": "IN_PUBLISHED_NDSE2_BIBLIOGRAPHY",
  "node_class": "FOWLER_SERIES_ADDITIVE_OVERLAY_V12",
  "ndse2_actual_chapters_json": "[\"C:C10\"]",
  "ndse2_in_bibliography": true,
  "ndse2_citation_basis": "Final 13.7.1 active EN/ITA source tree; index links and nocite are not chapter citations",
  "historical_v12_book_citation_status": "NOT_ADDED_TO_BOOK",
  "historical_v12_new_manuscript_citation": false
}

• → W101 · Articulate but Wrong: Self-Review Failures in LLM-Based Code Modernization
  F_EDITORIAL_COMPARISON · edge f00373 · key MF_V12:F_EDITORIAL_COMPARISON

These retain the graph’s exact relation labels. A connection is not automatically a causal claim or an independent replication.

{
  "label": "Guiding an LLM for Robust Java ByteBuffer Code",
  "kind": "practitioner_essay",
  "collection": "F",
  "authors": "Unmesh Joshi",
  "authors_json": "[\"Unmesh Joshi\"]",
  "publication_date": "2025-04-17",
  "date_provenance": "Dated entry on the supplied series index; article date also used where printed",
  "url": "https://martinfowler.com/articles/exploring-gen-ai/14-guiding-llm-java-bytebuffer.html",
  "publisher": "martinfowler.com",
  "series": "Exploring Generative AI",
  "series_url": "https://martinfowler.com/articles/exploring-gen-ai.html",
  "snapshot_date": "2026-10-13",
  "contribution": "Follows a developer–LLM dialogue about Java ByteBuffer header access. The engineer uncovers problems with mutable buffer position and guides the design toward clearer access and encapsulation.",
  "limitation": "An illustrative dialogue; removing position mutation does not by itself establish thread safety or general performance.",
  "evidence_class": "Worked programming example",
  "source_locator": "Body: header reading, position side effects and absolute access",
  "verification_scope": "Primary article byline and selected body passages inspected; no independent execution or statistical replication",
  "original_data_replicated": false,
  "new_manuscript_citation": false,
  "source_body_archived": false,
  "primary_stream": "T:Code quality & security",
  "streams_json": "[\"T:Code quality & security\", \"T:Verification & review\"]",
  "topics_json": "[\"K:Architecture and modularity\", \"K:Independent verification and oracles\", \"K:System understanding and rationale\"]",
  "editorial_chapters_json": "[\"C:C05\", \"C:C07\", \"C:C10\"]",
  "mapping_status": "Independent editorial mapping, not classification by the article authors or SE@PoliMi",
  "book_citation_status": "COMPANION_READING_NOT_IN_BOOK",
  "node_class": "FOWLER_SERIES_ADDITIVE_OVERLAY_V12",
  "ndse2_actual_chapters_json": "[]",
  "ndse2_in_bibliography": false,
  "ndse2_citation_basis": "Final 13.7.1 active EN/ITA source tree; index links and nocite are not chapter citations",
  "historical_v12_book_citation_status": "NOT_ADDED_TO_BOOK"
}

• → W118 · How AI Impacts Skill Formation
  F_EDITORIAL_COMPARISON · edge f00374 · key MF_V12:F_EDITORIAL_COMPARISON

These retain the graph’s exact relation labels. A connection is not automatically a causal claim or an independent replication.

{
  "label": "Building TMT Mirror Visualization with LLM: A Step-by-Step Journey",
  "kind": "practitioner_essay",
  "collection": "F",
  "authors": "Unmesh Joshi",
  "authors_json": "[\"Unmesh Joshi\"]",
  "publication_date": "2025-04-30",
  "date_provenance": "Dated entry on the supplied series index; article date also used where printed",
  "url": "https://martinfowler.com/articles/exploring-gen-ai/15-building-tmt-mirror-visualization.html",
  "publisher": "martinfowler.com",
  "series": "Exploring Generative AI",
  "series_url": "https://martinfowler.com/articles/exploring-gen-ai.html",
  "snapshot_date": "2026-10-13",
  "contribution": "Describes building a telescope-mirror visualization with unfamiliar tools. Small geometric steps, visual checks and domain knowledge guide successive changes while the developer learns the implementation approach.",
  "limitation": "A personal prototype-building account; neither retained learning nor broader productivity was evaluated.",
  "evidence_class": "Exploratory learning case",
  "source_locator": "Body: initial visualization, hexagon geometry and iterative refinement",
  "verification_scope": "Primary article byline and selected body passages inspected; no independent execution or statistical replication",
  "original_data_replicated": false,
  "new_manuscript_citation": false,
  "source_body_archived": false,
  "primary_stream": "T:Skills & learning",
  "streams_json": "[\"T:Skills & learning\", \"T:What agents can do\"]",
  "topics_json": "[\"K:Skills formation and teaching\", \"K:Human-agent collaboration\", \"K:Developer workflow and adoption\"]",
  "editorial_chapters_json": "[\"C:C03\", \"C:C04\", \"C:C10\"]",
  "mapping_status": "Independent editorial mapping, not classification by the article authors or SE@PoliMi",
  "book_citation_status": "COMPANION_READING_NOT_IN_BOOK",
  "node_class": "FOWLER_SERIES_ADDITIVE_OVERLAY_V12",
  "ndse2_actual_chapters_json": "[]",
  "ndse2_in_bibliography": false,
  "ndse2_citation_basis": "Final 13.7.1 active EN/ITA source tree; index links and nocite are not chapter citations",
  "historical_v12_book_citation_status": "NOT_ADDED_TO_BOOK"
}

• → W060 · Your AI, My Shell: Demystifying Prompt Injection Attacks on Agentic AI Coding Editors
  F_EDITORIAL_COMPARISON · edge f00375 · key MF_V12:F_EDITORIAL_COMPARISON
• → W053 · Scanning the Harness: Configuration Exposures in AI Coding-Agent Supply Chains
  F_EDITORIAL_COMPARISON · edge f00376 · key MF_V12:F_EDITORIAL_COMPARISON
• ← NDSE2:C06 · Securing Generated Code and Agent Runtimes
  NDSE2_CHAPTER_CITES · edge u20261018_00501 · key UPDATE_20261018:NDSE2_CHAPTER_CITES:EN
• ← NDSE2:C06 · Securing Generated Code and Agent Runtimes
  NDSE2_CHAPTER_CITES · edge u20261018_00511 · key UPDATE_20261018:NDSE2_CHAPTER_CITES:ITA

These retain the graph’s exact relation labels. A connection is not automatically a causal claim or an independent replication.

{
  "label": "Coding Assistants Threaten the Software Supply Chain",
  "kind": "practitioner_essay",
  "collection": "F",
  "authors": "Jim Gumbley; Lilly Ryan",
  "authors_json": "[\"Jim Gumbley\", \"Lilly Ryan\"]",
  "publication_date": "2025-05-13",
  "date_provenance": "Dated entry on the supplied series index; article date also used where printed",
  "url": "https://martinfowler.com/articles/exploring-gen-ai/software-supply-chain-attack-surface.html",
  "publisher": "martinfowler.com",
  "series": "Exploring Generative AI",
  "series_url": "https://martinfowler.com/articles/exploring-gen-ai.html",
  "snapshot_date": "2026-10-13",
  "contribution": "Analyzes the attack surface created when coding assistants consume project instructions and invoke tools with developer privileges. Discusses poisoned context, dependency risks and boundaries around execution.",
  "limitation": "Security analysis rather than an attack-incidence study. Protocol and tool details belong to the May 2025 publication context.",
  "evidence_class": "Threat analysis",
  "source_locator": "Body: assistant inputs, tools, supply-chain exposure and defenses",
  "verification_scope": "Primary article byline and selected body passages inspected; no independent execution or statistical replication",
  "original_data_replicated": false,
  "new_manuscript_citation": true,
  "source_body_archived": false,
  "primary_stream": "T:Code quality & security",
  "streams_json": "[\"T:Code quality & security\", \"T:Engineering structure\"]",
  "topics_json": "[\"K:Prompt injection and malicious tools\", \"K:Dependencies and agent supply chain\"]",
  "editorial_chapters_json": "[\"C:C03\", \"C:C06\"]",
  "mapping_status": "Independent editorial mapping, not classification by the article authors or SE@PoliMi",
  "book_citation_status": "IN_PUBLISHED_NDSE2_BIBLIOGRAPHY",
  "node_class": "FOWLER_SERIES_ADDITIVE_OVERLAY_V12",
  "ndse2_actual_chapters_json": "[\"C:C06\"]",
  "ndse2_in_bibliography": true,
  "ndse2_citation_basis": "Final 13.7.1 active EN/ITA source tree; index links and nocite are not chapter citations",
  "historical_v12_book_citation_status": "NOT_ADDED_TO_BOOK",
  "historical_v12_new_manuscript_citation": false
}

• → B022 · Static Context Engineering: Teaching AI About Your Project
  F_EDITORIAL_COMPARISON · edge f00377 · key MF_V12:F_EDITORIAL_COMPARISON
• ← F035 · When Software Meets Hardware
  F_EDITORIAL_COMPARISON · edge u20261018_00032 · key UPDATE_20261018:F_EDITORIAL_COMPARISON

These retain the graph’s exact relation labels. A connection is not automatically a causal claim or an independent replication.

{
  "label": "Building Custom Tooling with LLMs",
  "kind": "practitioner_essay",
  "collection": "F",
  "authors": "Unmesh Joshi",
  "authors_json": "[\"Unmesh Joshi\"]",
  "publication_date": "2025-05-14",
  "date_provenance": "Dated entry on the supplied series index; article date also used where printed",
  "url": "https://martinfowler.com/articles/exploring-gen-ai/16-building-custom-tooling-with-llms.html",
  "publisher": "martinfowler.com",
  "series": "Exploring Generative AI",
  "series_url": "https://martinfowler.com/articles/exploring-gen-ai.html",
  "snapshot_date": "2026-10-13",
  "contribution": "Recounts extending PlantUML with custom visualization tooling. The developer uses generation and feedback to build the tool, while shaping abstractions and keeping the resulting code understandable and maintainable.",
  "limitation": "One custom-tooling case; lower perceived effort is not a controlled productivity estimate.",
  "evidence_class": "Practitioner implementation case",
  "source_locator": "Body: PlantUML extension, implementation iterations and code habitability",
  "verification_scope": "Primary article byline and selected body passages inspected; no independent execution or statistical replication",
  "original_data_replicated": false,
  "new_manuscript_citation": false,
  "source_body_archived": false,
  "primary_stream": "T:Code quality & security",
  "streams_json": "[\"T:Code quality & security\", \"T:Productivity\", \"T:System knowledge & memory\"]",
  "topics_json": "[\"K:Architecture and modularity\", \"K:System understanding and rationale\", \"K:Developer workflow and adoption\"]",
  "editorial_chapters_json": "[\"C:C03\", \"C:C05\", \"C:C09\", \"C:C10\"]",
  "mapping_status": "Independent editorial mapping, not classification by the article authors or SE@PoliMi",
  "book_citation_status": "COMPANION_READING_NOT_IN_BOOK",
  "node_class": "FOWLER_SERIES_ADDITIVE_OVERLAY_V12",
  "ndse2_actual_chapters_json": "[]",
  "ndse2_in_bibliography": false,
  "ndse2_citation_basis": "Final 13.7.1 active EN/ITA source tree; index links and nocite are not chapter citations",
  "historical_v12_book_citation_status": "NOT_ADDED_TO_BOOK"
}

• ← F023 · Understanding Spec-Driven-Development: Kiro, spec-kit, and Tessl
  F_BODY_LINKS_TO · edge f00241 · key MF_V12:F_BODY_LINKS_TO
• → W111 · Many SWE-bench-Passing PRs Would Not Be Merged into Main
  F_EDITORIAL_COMPARISON · edge f00378 · key MF_V12:F_EDITORIAL_COMPARISON
• → W031 · More Code, Less Reuse: Investigating Code Quality and Reviewer Sentiment towards AI-generated Pull Requests
  F_EDITORIAL_COMPARISON · edge f00379 · key MF_V12:F_EDITORIAL_COMPARISON

These retain the graph’s exact relation labels. A connection is not automatically a causal claim or an independent replication.

{
  "label": "Autonomous coding agents: A Codex example",
  "kind": "practitioner_essay",
  "collection": "F",
  "authors": "Birgitta Böckeler",
  "authors_json": "[\"Birgitta Böckeler\"]",
  "publication_date": "2025-06-04",
  "date_provenance": "Dated entry on the supplied series index; article date also used where printed",
  "url": "https://martinfowler.com/articles/exploring-gen-ai/autonomous-agents-codex-example.html",
  "publisher": "martinfowler.com",
  "series": "Exploring Generative AI",
  "series_url": "https://martinfowler.com/articles/exploring-gen-ai.html",
  "snapshot_date": "2026-10-13",
  "contribution": "Inspects a background agent making a small Haiven change. The execution trace reveals code-search behavior, missing executable checks and inconsistent reuse across attempts, distinguishing unattended execution from accepted work.",
  "limitation": "A small dated tool experiment; not a model ranking or a general estimate of autonomous success.",
  "evidence_class": "Exploratory agent case",
  "source_locator": "Body: supervised/background distinction, task trace and repeated attempts",
  "verification_scope": "Primary article byline and selected body passages inspected; no independent execution or statistical replication",
  "original_data_replicated": false,
  "new_manuscript_citation": false,
  "source_body_archived": false,
  "primary_stream": "T:What agents can do",
  "streams_json": "[\"T:What agents can do\", \"T:Verification & review\"]",
  "topics_json": "[\"K:Agent failure and escalation\", \"K:Code reuse and duplication\", \"K:Independent verification and oracles\"]",
  "editorial_chapters_json": "[\"C:C03\", \"C:C05\", \"C:C07\", \"C:C09\"]",
  "mapping_status": "Independent editorial mapping, not classification by the article authors or SE@PoliMi",
  "book_citation_status": "COMPANION_READING_NOT_IN_BOOK",
  "node_class": "FOWLER_SERIES_ADDITIVE_OVERLAY_V12",
  "ndse2_actual_chapters_json": "[]",
  "ndse2_in_bibliography": false,
  "ndse2_citation_basis": "Final 13.7.1 active EN/ITA source tree; index links and nocite are not chapter citations",
  "historical_v12_book_citation_status": "NOT_ADDED_TO_BOOK"
}

• → F007 · How is GenAI different from other code generators?
  F_BODY_LINKS_TO · edge f00197 · key MF_V12:F_BODY_LINKS_TO
• → W189 · The Substrate Collapse: AI Code Generation Invalidates Authorship-Based Knowledge Metrics
  F_EDITORIAL_COMPARISON · edge f00380 · key MF_V12:F_EDITORIAL_COMPARISON

These retain the graph’s exact relation labels. A connection is not automatically a causal claim or an independent replication.

{
  "label": "I still care about the code",
  "kind": "practitioner_essay",
  "collection": "F",
  "authors": "Birgitta Böckeler",
  "authors_json": "[\"Birgitta Böckeler\"]",
  "publication_date": "2025-07-09",
  "date_provenance": "Dated entry on the supplied series index; article date also used where printed",
  "url": "https://martinfowler.com/articles/exploring-gen-ai/i-still-care-about-the-code.html",
  "publisher": "martinfowler.com",
  "series": "Exploring Generative AI",
  "series_url": "https://martinfowler.com/articles/exploring-gen-ai.html",
  "snapshot_date": "2026-10-13",
  "contribution": "Argues that responsibility for operational software still requires understanding code and its risks. Contrasts probabilistic generation with designed language translation and relates oversight to possible error consequences.",
  "limitation": "A professional position, not a calibrated threshold for how much review every change requires.",
  "evidence_class": "Conceptual argument",
  "source_locator": "Body: operational responsibility and comparison with compilers",
  "verification_scope": "Primary article byline and selected body passages inspected; no independent execution or statistical replication",
  "original_data_replicated": false,
  "new_manuscript_citation": false,
  "source_body_archived": false,
  "primary_stream": "T:Verification & review",
  "streams_json": "[\"T:Verification & review\", \"T:Code quality & security\"]",
  "topics_json": "[\"K:Automation bias and trust\", \"K:System understanding and rationale\", \"K:Ownership and knowledge transfer\"]",
  "editorial_chapters_json": "[\"C:C01\", \"C:C05\", \"C:C08\", \"C:C13\"]",
  "mapping_status": "Independent editorial mapping, not classification by the article authors or SE@PoliMi",
  "book_citation_status": "COMPANION_READING_NOT_IN_BOOK",
  "node_class": "FOWLER_SERIES_ADDITIVE_OVERLAY_V12",
  "ndse2_actual_chapters_json": "[]",
  "ndse2_in_bibliography": false,
  "ndse2_citation_basis": "Final 13.7.1 active EN/ITA source tree; index links and nocite are not chapter citations",
  "historical_v12_book_citation_status": "NOT_ADDED_TO_BOOK"
}

• → W009 · Productivity Assessment of Neural Code Completion
  F_EDITORIAL_COMPARISON · edge f00381 · key MF_V12:F_EDITORIAL_COMPARISON
• ← F034 · Practitioner Voice
  F_EDITORIAL_COMPARISON · edge u20261018_00030 · key UPDATE_20261018:F_EDITORIAL_COMPARISON

These retain the graph’s exact relation labels. A connection is not automatically a causal claim or an independent replication.

{
  "label": "Partner with the AI, throw away the code",
  "kind": "practitioner_essay",
  "collection": "F",
  "authors": "Matteo Vaccari",
  "authors_json": "[\"Matteo Vaccari\"]",
  "publication_date": "2025-07-31",
  "date_provenance": "Dated entry on the supplied series index; article date also used where printed",
  "url": "https://martinfowler.com/articles/exploring-gen-ai/partner-with-ai-and-throw-away-the-code.html",
  "publisher": "martinfowler.com",
  "series": "Exploring Generative AI",
  "series_url": "https://martinfowler.com/articles/exploring-gen-ai.html",
  "snapshot_date": "2026-10-13",
  "contribution": "In a Go/MySQL performance investigation, generated helpers, experiments and unsuccessful rewrites helped the developer understand the problem before rewriting the core solution. Useful assistance can therefore leave little accepted generated code.",
  "limitation": "A single client case; code-acceptance rate is questioned, but a general AI-caused productivity gain is not measured.",
  "evidence_class": "Practitioner case report",
  "source_locator": "Body: diagnosis, experiments, discarded proposals and final rewrite",
  "verification_scope": "Primary article byline and selected body passages inspected; no independent execution or statistical replication",
  "original_data_replicated": false,
  "new_manuscript_citation": false,
  "source_body_archived": false,
  "primary_stream": "T:Productivity",
  "streams_json": "[\"T:Productivity\", \"T:Skills & learning\", \"T:Economics\"]",
  "topics_json": "[\"K:Output versus shipped value\", \"K:Evaluation methodology and baselines\", \"K:System understanding and rationale\"]",
  "editorial_chapters_json": "[\"C:C02\", \"C:C05\", \"C:C10\", \"C:C12\"]",
  "mapping_status": "Independent editorial mapping, not classification by the article authors or SE@PoliMi",
  "book_citation_status": "COMPANION_READING_NOT_IN_BOOK",
  "node_class": "FOWLER_SERIES_ADDITIVE_OVERLAY_V12",
  "ndse2_actual_chapters_json": "[]",
  "ndse2_in_bibliography": false,
  "ndse2_citation_basis": "Final 13.7.1 active EN/ITA source tree; index links and nocite are not chapter citations",
  "historical_v12_book_citation_status": "NOT_ADDED_TO_BOOK"
}

• → W093 · Automating Low-Risk Code Review at Meta: RADAR, Risk Calibration, and Review Efficiency
  F_EDITORIAL_COMPARISON · edge f00382 · key MF_V12:F_EDITORIAL_COMPARISON
• → P118 · What is happening with code reviews?
  F_EDITORIAL_COMPARISON · edge f00383 · key MF_V12:F_EDITORIAL_COMPARISON

These retain the graph’s exact relation labels. A connection is not automatically a causal claim or an independent replication.

{
  "label": "To vibe or not to vibe",
  "kind": "practitioner_essay",
  "collection": "F",
  "authors": "Birgitta Böckeler",
  "authors_json": "[\"Birgitta Böckeler\"]",
  "publication_date": "2025-09-23",
  "date_provenance": "Dated entry on the supplied series index; article date also used where printed",
  "url": "https://martinfowler.com/articles/exploring-gen-ai/to-vibe-or-not-vibe.html",
  "publisher": "martinfowler.com",
  "series": "Exploring Generative AI",
  "series_url": "https://martinfowler.com/articles/exploring-gen-ai.html",
  "snapshot_date": "2026-10-13",
  "contribution": "Describes decisions about oversight using the chance of an error, its impact and the ability to detect it. The examples favor situation-dependent attention over a single rule for all generated code.",
  "limitation": "A qualitative risk heuristic and personal experience; not a statistically calibrated risk model.",
  "evidence_class": "Practitioner decision heuristic",
  "source_locator": "Body: risk factors and situational review decisions",
  "verification_scope": "Primary article byline and selected body passages inspected; no independent execution or statistical replication",
  "original_data_replicated": false,
  "new_manuscript_citation": false,
  "source_body_archived": false,
  "primary_stream": "T:Verification & review",
  "streams_json": "[\"T:Verification & review\", \"T:Engineering structure\"]",
  "topics_json": "[\"K:Automation bias and trust\", \"K:Human-agent collaboration\"]",
  "editorial_chapters_json": "[\"C:C01\", \"C:C08\", \"C:C13\"]",
  "mapping_status": "Independent editorial mapping, not classification by the article authors or SE@PoliMi",
  "book_citation_status": "COMPANION_READING_NOT_IN_BOOK",
  "node_class": "FOWLER_SERIES_ADDITIVE_OVERLAY_V12",
  "ndse2_actual_chapters_json": "[]",
  "ndse2_in_bibliography": false,
  "ndse2_citation_basis": "Final 13.7.1 active EN/ITA source tree; index links and nocite are not chapter citations",
  "historical_v12_book_citation_status": "NOT_ADDED_TO_BOOK"
}

• → B022 · Static Context Engineering: Teaching AI About Your Project
  F_EDITORIAL_COMPARISON · edge f00384 · key MF_V12:F_EDITORIAL_COMPARISON

These retain the graph’s exact relation labels. A connection is not automatically a causal claim or an independent replication.

{
  "label": "Anchoring AI to a reference application",
  "kind": "practitioner_essay",
  "collection": "F",
  "authors": "Birgitta Böckeler",
  "authors_json": "[\"Birgitta Böckeler\"]",
  "publication_date": "2025-09-25",
  "date_provenance": "Dated entry on the supplied series index; article date also used where printed",
  "url": "https://martinfowler.com/articles/exploring-gen-ai/anchoring-to-reference.html",
  "publisher": "martinfowler.com",
  "series": "Exploring Generative AI",
  "series_url": "https://martinfowler.com/articles/exploring-gen-ai.html",
  "snapshot_date": "2026-10-13",
  "contribution": "Explores using a runnable reference application to guide changes across services. A scoped comparison and a reviewed change report precede implementation, raising a choice between AI-based propagation and deterministic transformations.",
  "limitation": "A small reference-update example; consistency and maintainability at organizational scale remain untested.",
  "evidence_class": "Exploratory workflow example",
  "source_locator": "Body: reference application, differences, report and update workflow",
  "verification_scope": "Primary article byline and selected body passages inspected; no independent execution or statistical replication",
  "original_data_replicated": false,
  "new_manuscript_citation": false,
  "source_body_archived": false,
  "primary_stream": "T:System knowledge & memory",
  "streams_json": "[\"T:System knowledge & memory\", \"T:Code quality & security\", \"T:Engineering structure\"]",
  "topics_json": "[\"K:Architecture and modularity\", \"K:Agent context and persistent memory\", \"K:Specifications and governance\"]",
  "editorial_chapters_json": "[\"C:C04\", \"C:C05\", \"C:C09\"]",
  "mapping_status": "Independent editorial mapping, not classification by the article authors or SE@PoliMi",
  "book_citation_status": "COMPANION_READING_NOT_IN_BOOK",
  "node_class": "FOWLER_SERIES_ADDITIVE_OVERLAY_V12",
  "ndse2_actual_chapters_json": "[]",
  "ndse2_in_bibliography": false,
  "ndse2_citation_basis": "Final 13.7.1 active EN/ITA source tree; index links and nocite are not chapter citations",
  "historical_v12_book_citation_status": "NOT_ADDED_TO_BOOK"
}

• → F018 · Autonomous coding agents: A Codex example
  F_BODY_LINKS_TO · edge f00241 · key MF_V12:F_BODY_LINKS_TO
• → O018 · How to write a good spec for AI agents
  F_EDITORIAL_COMPARISON · edge f00385 · key MF_V12:F_EDITORIAL_COMPARISON
• → B063 · Advanced Patterns: From Frameworks to Practice
  F_EDITORIAL_COMPARISON · edge f00386 · key MF_V12:F_EDITORIAL_COMPARISON
• → B064 · Limitations and Professional Judgment
  F_EDITORIAL_COMPARISON · edge f00387 · key MF_V12:F_EDITORIAL_COMPARISON
• ← NDSE2:C04 · Intentions, Contracts and Iterative Change
  NDSE2_CHAPTER_CITES · edge u20261018_00315 · key UPDATE_20261018:NDSE2_CHAPTER_CITES:EN
• ← NDSE2:C04 · Intentions, Contracts and Iterative Change
  NDSE2_CHAPTER_CITES · edge u20261018_00330 · key UPDATE_20261018:NDSE2_CHAPTER_CITES:ITA

These retain the graph’s exact relation labels. A connection is not automatically a causal claim or an independent replication.

{
  "label": "Understanding Spec-Driven-Development: Kiro, spec-kit, and Tessl",
  "kind": "practitioner_essay",
  "collection": "F",
  "authors": "Birgitta Böckeler",
  "authors_json": "[\"Birgitta Böckeler\"]",
  "publication_date": "2025-10-15",
  "date_provenance": "Dated entry on the supplied series index; article date also used where printed",
  "url": "https://martinfowler.com/articles/exploring-gen-ai/sdd-3-tools.html",
  "publisher": "martinfowler.com",
  "series": "Exploring Generative AI",
  "series_url": "https://martinfowler.com/articles/exploring-gen-ai.html",
  "snapshot_date": "2026-10-13",
  "contribution": "Distinguishes spec-first, spec-anchored and spec-as-source approaches through three tool explorations. Examines persistent specifications, task sizing, brownfield fit, instruction following and documentation overhead.",
  "limitation": "A dated exploratory comparison, not controlled evidence for SDD effectiveness. The author’s questions about specification overhead and source-of-truth claims are retained.",
  "evidence_class": "Practitioner comparative exploration",
  "source_locator": "Body: definitions of SDD; Kiro, spec-kit and Tessl explorations; reflections",
  "verification_scope": "Primary article byline and selected body passages inspected; no independent execution or statistical replication",
  "original_data_replicated": false,
  "new_manuscript_citation": true,
  "source_body_archived": false,
  "primary_stream": "T:Engineering structure",
  "streams_json": "[\"T:Engineering structure\", \"T:System knowledge & memory\"]",
  "topics_json": "[\"K:Specifications and governance\", \"K:Developer workflow and adoption\", \"K:Agent context and persistent memory\"]",
  "editorial_chapters_json": "[\"C:C04\", \"C:C09\", \"C:C11\", \"C:C13\"]",
  "mapping_status": "Independent editorial mapping, not classification by the article authors or SE@PoliMi",
  "book_citation_status": "IN_PUBLISHED_NDSE2_BIBLIOGRAPHY",
  "node_class": "FOWLER_SERIES_ADDITIVE_OVERLAY_V12",
  "ndse2_actual_chapters_json": "[\"C:C04\"]",
  "ndse2_in_bibliography": true,
  "ndse2_citation_basis": "Final 13.7.1 active EN/ITA source tree; index links and nocite are not chapter citations",
  "historical_v12_book_citation_status": "NOT_ADDED_TO_BOOK",
  "historical_v12_new_manuscript_citation": false
}

• → O007 · Agentic Code Quality
  F_EDITORIAL_COMPARISON · edge f00388 · key MF_V12:F_EDITORIAL_COMPARISON
• ← NDSE2:C05 · Quality, Architecture and Maintainability
  NDSE2_CHAPTER_CITES · edge u20261018_00399 · key UPDATE_20261018:NDSE2_CHAPTER_CITES:EN
• ← NDSE2:C05 · Quality, Architecture and Maintainability
  NDSE2_CHAPTER_CITES · edge u20261018_00417 · key UPDATE_20261018:NDSE2_CHAPTER_CITES:ITA

These retain the graph’s exact relation labels. A connection is not automatically a causal claim or an independent replication.

{
  "label": "Assessing internal quality while coding with an agent",
  "kind": "practitioner_essay",
  "collection": "F",
  "authors": "Erik Doernenburg",
  "authors_json": "[\"Erik Doernenburg\"]",
  "publication_date": "2026-01-27",
  "date_provenance": "Dated entry on the supplied series index; article date also used where printed",
  "url": "https://martinfowler.com/articles/exploring-gen-ai/ccmenu-quality.html",
  "publisher": "martinfowler.com",
  "series": "Exploring Generative AI",
  "series_url": "https://martinfowler.com/articles/exploring-gen-ai.html",
  "snapshot_date": "2026-10-13",
  "contribution": "Uses a CCMenu feature change to inspect internal quality. A fix that compiles can still erode the intended type model, introduce duplication or add unnecessary mechanisms; expert examination exposes those differences.",
  "limitation": "The author presents an anecdote, not a study; observations do not establish a population-level quality effect.",
  "evidence_class": "Practitioner code-review case",
  "source_locator": "Body: GitLab support, optional values, duplication and cache behavior",
  "verification_scope": "Primary article byline and selected body passages inspected; no independent execution or statistical replication",
  "original_data_replicated": false,
  "new_manuscript_citation": true,
  "source_body_archived": false,
  "primary_stream": "T:Code quality & security",
  "streams_json": "[\"T:Code quality & security\", \"T:Verification & review\"]",
  "topics_json": "[\"K:Technical debt and complexity\", \"K:Code reuse and duplication\", \"K:Architecture and modularity\"]",
  "editorial_chapters_json": "[\"C:C05\", \"C:C07\", \"C:C08\"]",
  "mapping_status": "Independent editorial mapping, not classification by the article authors or SE@PoliMi",
  "book_citation_status": "IN_PUBLISHED_NDSE2_BIBLIOGRAPHY",
  "node_class": "FOWLER_SERIES_ADDITIVE_OVERLAY_V12",
  "ndse2_actual_chapters_json": "[\"C:C05\"]",
  "ndse2_in_bibliography": true,
  "ndse2_citation_basis": "Final 13.7.1 active EN/ITA source tree; index links and nocite are not chapter citations",
  "historical_v12_book_citation_status": "NOT_ADDED_TO_BOOK",
  "historical_v12_new_manuscript_citation": false
}

• → O031 · Context Engineering: Bringing Engineering Discipline to Prompts
  F_EDITORIAL_COMPARISON · edge f00389 · key MF_V12:F_EDITORIAL_COMPARISON
• → B090 · Memory and Context Engineering
  F_EDITORIAL_COMPARISON · edge f00390 · key MF_V12:F_EDITORIAL_COMPARISON
• ← NDSE2:C09 · System Understanding, Context and Persistent Memory
  NDSE2_CHAPTER_CITES · edge u20261018_00794 · key UPDATE_20261018:NDSE2_CHAPTER_CITES:EN
• ← NDSE2:C09 · System Understanding, Context and Persistent Memory
  NDSE2_CHAPTER_CITES · edge u20261018_00811 · key UPDATE_20261018:NDSE2_CHAPTER_CITES:ITA

These retain the graph’s exact relation labels. A connection is not automatically a causal claim or an independent replication.

{
  "label": "Context Engineering for Coding Agents",
  "kind": "practitioner_essay",
  "collection": "F",
  "authors": "Birgitta Böckeler",
  "authors_json": "[\"Birgitta Böckeler\"]",
  "publication_date": "2026-02-05",
  "date_provenance": "Dated entry on the supplied series index; article date also used where printed",
  "url": "https://martinfowler.com/articles/exploring-gen-ai/context-engineering-coding-agents.html",
  "publisher": "martinfowler.com",
  "series": "Exploring Generative AI",
  "series_url": "https://martinfowler.com/articles/exploring-gen-ai.html",
  "snapshot_date": "2026-10-13",
  "contribution": "Organizes context features by who invokes them: a person, the model or agent software. Discusses rules, skills, subagents, external context and hooks as parts of the developer experience.",
  "limitation": "Feature and practice snapshot, not proof of effectiveness or enforcement. Adding more configuration can produce contradictions or an illusion of control.",
  "evidence_class": "Practitioner technical taxonomy",
  "source_locator": "Body: trigger-based feature overview and context configuration",
  "verification_scope": "Primary article byline and selected body passages inspected; no independent execution or statistical replication",
  "original_data_replicated": false,
  "new_manuscript_citation": true,
  "source_body_archived": false,
  "primary_stream": "T:System knowledge & memory",
  "streams_json": "[\"T:System knowledge & memory\", \"T:Engineering structure\"]",
  "topics_json": "[\"K:Agent context and persistent memory\", \"K:Developer workflow and adoption\", \"K:Specifications and governance\"]",
  "editorial_chapters_json": "[\"C:C03\", \"C:C09\", \"C:C13\"]",
  "mapping_status": "Independent editorial mapping, not classification by the article authors or SE@PoliMi",
  "book_citation_status": "IN_PUBLISHED_NDSE2_BIBLIOGRAPHY",
  "node_class": "FOWLER_SERIES_ADDITIVE_OVERLAY_V12",
  "ndse2_actual_chapters_json": "[\"C:C09\"]",
  "ndse2_in_bibliography": true,
  "ndse2_citation_basis": "Final 13.7.1 active EN/ITA source tree; index links and nocite are not chapter citations",
  "historical_v12_book_citation_status": "NOT_ADDED_TO_BOOK",
  "historical_v12_new_manuscript_citation": false
}

• → B104 · Non-Deterministic Loop Engineering: A Critical Assessment
  F_EDITORIAL_COMPARISON · edge f00391 · key MF_V12:F_EDITORIAL_COMPARISON
• ← NDSE2:C13 · Choosing Practices and Recognizing Their Limits
  NDSE2_CHAPTER_CITES · edge u20261018_01184 · key UPDATE_20261018:NDSE2_CHAPTER_CITES:EN
• ← NDSE2:C13 · Choosing Practices and Recognizing Their Limits
  NDSE2_CHAPTER_CITES · edge u20261018_01190 · key UPDATE_20261018:NDSE2_CHAPTER_CITES:ITA

These retain the graph’s exact relation labels. A connection is not automatically a causal claim or an independent replication.

{
  "label": "Harness Engineering - first thoughts",
  "kind": "practitioner_essay",
  "collection": "F",
  "authors": "Birgitta Böckeler",
  "authors_json": "[\"Birgitta Böckeler\"]",
  "publication_date": "2026-02-17",
  "date_provenance": "Dated entry on the supplied series index; article date also used where printed",
  "url": "https://martinfowler.com/articles/exploring-gen-ai/harness-engineering-memo.html",
  "publisher": "martinfowler.com",
  "series": "Exploring Generative AI",
  "series_url": "https://martinfowler.com/articles/exploring-gen-ai.html",
  "snapshot_date": "2026-10-13",
  "contribution": "Interprets a reported harness through context, architectural constraints and codebase maintenance. Raises questions about reusable harnesses, constrained topologies and the effort of retrofitting existing systems.",
  "limitation": "A reaction to another organization’s account, not an independent experiment. The author explicitly notes a gap in the reported functional verification.",
  "evidence_class": "Interpretive practitioner commentary",
  "source_locator": "Body: three categories; templates; runtime constraints; brownfield retrofit; final thoughts",
  "verification_scope": "Primary article byline and selected body passages inspected; no independent execution or statistical replication",
  "original_data_replicated": false,
  "new_manuscript_citation": true,
  "source_body_archived": false,
  "primary_stream": "T:Engineering structure",
  "streams_json": "[\"T:Engineering structure\", \"T:Code quality & security\"]",
  "topics_json": "[\"K:Architecture and modularity\", \"K:Independent verification and oracles\", \"K:Long-running agent execution\"]",
  "editorial_chapters_json": "[\"C:C03\", \"C:C05\", \"C:C07\", \"C:C11\", \"C:C13\"]",
  "mapping_status": "Independent editorial mapping, not classification by the article authors or SE@PoliMi",
  "book_citation_status": "IN_PUBLISHED_NDSE2_BIBLIOGRAPHY",
  "node_class": "FOWLER_SERIES_ADDITIVE_OVERLAY_V12",
  "source_version_status": "EARLIER_MEMO_SUPERSEDED_BY_EXPANDED_ARTICLE",
  "latest_version_url": "https://martinfowler.com/articles/harness-engineering.html",
  "latest_version_title": "Harness engineering for coding agent users",
  "latest_version_publication_date": "2026-04-02",
  "supersession_verified_date": "2026-10-18",
  "source_version_note": "The later primary article explicitly supersedes this 17 February memo. Direct retrieval still exposes the historical memo with a pointer to the expansion; an HTTP redirect was not established.",
  "ndse2_actual_chapters_json": "[\"C:C13\"]",
  "ndse2_in_bibliography": true,
  "ndse2_citation_basis": "Final 13.7.1 active EN/ITA source tree; index links and nocite are not chapter citations",
  "historical_v12_book_citation_status": "NOT_ADDED_TO_BOOK",
  "historical_v12_new_manuscript_citation": false
}

• → O009 · Own the Outer Loop
  F_EDITORIAL_COMPARISON · edge f00392 · key MF_V12:F_EDITORIAL_COMPARISON
• → B105 · The Organization Becomes a Loop
  F_EDITORIAL_COMPARISON · edge f00393 · key MF_V12:F_EDITORIAL_COMPARISON
• ← NDSE2:C11 · Teams, Governance and the Engineering Organization
  NDSE2_CHAPTER_CITES · edge u20261018_00984 · key UPDATE_20261018:NDSE2_CHAPTER_CITES:EN
• ← NDSE2:C11 · Teams, Governance and the Engineering Organization
  NDSE2_CHAPTER_CITES · edge u20261018_01003 · key UPDATE_20261018:NDSE2_CHAPTER_CITES:ITA

These retain the graph’s exact relation labels. A connection is not automatically a causal claim or an independent replication.

{
  "label": "Humans and Agents in Software Engineering Loops",
  "kind": "practitioner_essay",
  "collection": "F",
  "authors": "Kief Morris",
  "authors_json": "[\"Kief Morris\"]",
  "publication_date": "2026-03-04",
  "date_provenance": "Dated entry on the supplied series index; article date also used where printed",
  "url": "https://martinfowler.com/articles/exploring-gen-ai/humans-and-agents.html",
  "publisher": "martinfowler.com",
  "series": "Exploring Generative AI",
  "series_url": "https://martinfowler.com/articles/exploring-gen-ai.html",
  "snapshot_date": "2026-10-13",
  "contribution": "Distinguishes the why loop that pursues outcomes from nested how loops that build software. Proposes human work on the loop: improving guidance, checks and workflow rather than only editing individual outputs.",
  "limitation": "A conceptual operating model. Its proposed agentic improvement flywheel is not a demonstrated organizational or reliability result.",
  "evidence_class": "Conceptual operating model",
  "source_locator": "Body: why/how loops; humans outside, in and on the loop; flywheel",
  "verification_scope": "Primary article byline and selected body passages inspected; no independent execution or statistical replication",
  "original_data_replicated": false,
  "new_manuscript_citation": true,
  "source_body_archived": false,
  "primary_stream": "T:Engineering structure",
  "streams_json": "[\"T:Engineering structure\", \"T:Jobs & roles\"]",
  "topics_json": "[\"K:Human-agent collaboration\", \"K:Long-running agent execution\", \"K:Specifications and governance\"]",
  "editorial_chapters_json": "[\"C:C01\", \"C:C03\", \"C:C11\", \"C:C13\"]",
  "mapping_status": "Independent editorial mapping, not classification by the article authors or SE@PoliMi",
  "book_citation_status": "IN_PUBLISHED_NDSE2_BIBLIOGRAPHY",
  "node_class": "FOWLER_SERIES_ADDITIVE_OVERLAY_V12",
  "ndse2_actual_chapters_json": "[\"C:C11\"]",
  "ndse2_in_bibliography": true,
  "ndse2_citation_basis": "Final 13.7.1 active EN/ITA source tree; index links and nocite are not chapter citations",
  "historical_v12_book_citation_status": "NOT_ADDED_TO_BOOK",
  "historical_v12_new_manuscript_citation": false
}

• → F029 · Experiences with local models for coding
  F_BODY_LINKS_TO · edge f00296 · key MF_V12:F_BODY_LINKS_TO
• ← F029 · Experiences with local models for coding
  F_BODY_LINKS_TO · edge f00309 · key MF_V12:F_BODY_LINKS_TO
• → B108 · Building a Full Local AI Loop
  F_EDITORIAL_COMPARISON · edge f00394 · key MF_V12:F_EDITORIAL_COMPARISON

These retain the graph’s exact relation labels. A connection is not automatically a causal claim or an independent replication.

{
  "label": "Viability of local models for coding",
  "kind": "practitioner_essay",
  "collection": "F",
  "authors": "Birgitta Böckeler",
  "authors_json": "[\"Birgitta Böckeler\"]",
  "publication_date": "2026-07-07",
  "date_provenance": "Dated entry on the supplied series index; article date also used where printed",
  "url": "https://martinfowler.com/articles/exploring-gen-ai/local-models-for-coding-factors.html",
  "publisher": "martinfowler.com",
  "series": "Exploring Generative AI",
  "series_url": "https://martinfowler.com/articles/exploring-gen-ai.html",
  "snapshot_date": "2026-10-13",
  "contribution": "Frames local agentic coding as a combined hardware, model, runtime, context and tool-calling problem. The account separates being able to run a model from obtaining usable results.",
  "limitation": "A personal hardware/tool snapshot. Explanations for observed configuration differences are hypotheses; code quality was not systematically evaluated.",
  "evidence_class": "Practitioner configuration exploration",
  "source_locator": "Scope; factors influencing viability; context size and harness; conclusions",
  "verification_scope": "Primary article byline and selected body passages inspected; no independent execution or statistical replication",
  "original_data_replicated": false,
  "new_manuscript_citation": false,
  "source_body_archived": false,
  "primary_stream": "T:What agents can do",
  "streams_json": "[\"T:What agents can do\", \"T:Economics\", \"T:System knowledge & memory\"]",
  "topics_json": "[\"K:Agent context and persistent memory\", \"K:Economic costs and return\", \"K:Evaluation methodology and baselines\"]",
  "editorial_chapters_json": "[\"C:C03\", \"C:C09\", \"C:C12\"]",
  "mapping_status": "Independent editorial mapping, not classification by the article authors or SE@PoliMi",
  "book_citation_status": "COMPANION_READING_NOT_IN_BOOK",
  "node_class": "FOWLER_SERIES_ADDITIVE_OVERLAY_V12",
  "ndse2_actual_chapters_json": "[]",
  "ndse2_in_bibliography": false,
  "ndse2_citation_basis": "Final 13.7.1 active EN/ITA source tree; index links and nocite are not chapter citations",
  "historical_v12_book_citation_status": "NOT_ADDED_TO_BOOK"
}

• ← F028 · Viability of local models for coding
  F_BODY_LINKS_TO · edge f00296 · key MF_V12:F_BODY_LINKS_TO
• → F028 · Viability of local models for coding
  F_BODY_LINKS_TO · edge f00309 · key MF_V12:F_BODY_LINKS_TO
• → F011 · Expanding the solution size with multi-file editing
  F_BODY_LINKS_TO · edge f00310 · key MF_V12:F_BODY_LINKS_TO
• → B109 · Choosing the Right Loop
  F_EDITORIAL_COMPARISON · edge f00395 · key MF_V12:F_EDITORIAL_COMPARISON

These retain the graph’s exact relation labels. A connection is not automatically a causal claim or an independent replication.

{
  "label": "Experiences with local models for coding",
  "kind": "practitioner_essay",
  "collection": "F",
  "authors": "Birgitta Böckeler",
  "authors_json": "[\"Birgitta Böckeler\"]",
  "publication_date": "2026-07-08",
  "date_provenance": "Dated entry on the supplied series index; article date also used where printed",
  "url": "https://martinfowler.com/articles/exploring-gen-ai/local-models-for-coding-experiences.html",
  "publisher": "martinfowler.com",
  "series": "Exploring Generative AI",
  "series_url": "https://martinfowler.com/articles/exploring-gen-ai.html",
  "snapshot_date": "2026-10-13",
  "contribution": "Presents a viability funnel and compares manual trials, a small automated evaluation and everyday local-model use. Outcome sensitivity to task, context and configuration makes simple capability claims difficult.",
  "limitation": "Two principal JavaScript/TypeScript tasks and small uneven repeat counts; one-shot automated trials differ from interactive work and do not establish a general ranking.",
  "evidence_class": "Exploratory task evaluation",
  "source_locator": "Viability funnel; two task descriptions; manual/automated phases; day-to-day use",
  "verification_scope": "Primary article byline and selected body passages inspected; no independent execution or statistical replication",
  "original_data_replicated": false,
  "new_manuscript_citation": false,
  "source_body_archived": false,
  "primary_stream": "T:What agents can do",
  "streams_json": "[\"T:What agents can do\", \"T:Productivity\", \"T:System knowledge & memory\"]",
  "topics_json": "[\"K:Benchmarks and construct validity\", \"K:Evaluation methodology and baselines\", \"K:Agent context and persistent memory\"]",
  "editorial_chapters_json": "[\"C:C02\", \"C:C03\", \"C:C09\", \"C:C12\"]",
  "mapping_status": "Independent editorial mapping, not classification by the article authors or SE@PoliMi",
  "book_citation_status": "COMPANION_READING_NOT_IN_BOOK",
  "node_class": "FOWLER_SERIES_ADDITIVE_OVERLAY_V12",
  "ndse2_actual_chapters_json": "[]",
  "ndse2_in_bibliography": false,
  "ndse2_citation_basis": "Final 13.7.1 active EN/ITA source tree; index links and nocite are not chapter citations",
  "historical_v12_book_citation_status": "NOT_ADDED_TO_BOOK"
}

• ← F031 · TDD inside the agent loop - theater or actual value?
  F_BODY_LINKS_TO · edge f00332 · key MF_V12:F_BODY_LINKS_TO
• ← F033 · Are Harnesses Worth the Cost?
  F_BODY_LINKS_TO · edge f00357 · key MF_V12:F_BODY_LINKS_TO
• → B106 · The Economics of Loop Engineering
  F_EDITORIAL_COMPARISON · edge f00396 · key MF_V12:F_EDITORIAL_COMPARISON

These retain the graph’s exact relation labels. A connection is not automatically a causal claim or an independent replication.

{
  "label": "The Economic Benefit of Refactoring",
  "kind": "practitioner_essay",
  "collection": "F",
  "authors": "Giles Edwards-Alexander",
  "authors_json": "[\"Giles Edwards-Alexander\"]",
  "publication_date": "2026-07-30",
  "date_provenance": "Dated entry on the supplied series index; article date also used where printed",
  "url": "https://martinfowler.com/articles/exploring-gen-ai/refactoring-economic-benefit.html",
  "publisher": "martinfowler.com",
  "series": "Exploring Generative AI",
  "series_url": "https://martinfowler.com/articles/exploring-gen-ai.html",
  "snapshot_date": "2026-10-13",
  "contribution": "Repeats one representative change as a large Rust data-access module is refactored. Estimated input-token use falls after decomposition permits more selective code reading, connecting architecture to inference work.",
  "limitation": "One application and change, approximate token counting and incomplete refactoring-cost accounting; this does not establish general ROI or savings for every future task.",
  "evidence_class": "Exploratory refactoring experiment",
  "source_locator": "Experiment in refactoring; results table; notes on process; further work",
  "verification_scope": "Primary article byline and selected body passages inspected; no independent execution or statistical replication",
  "original_data_replicated": false,
  "new_manuscript_citation": false,
  "source_body_archived": false,
  "primary_stream": "T:Economics",
  "streams_json": "[\"T:Economics\", \"T:Code quality & security\"]",
  "topics_json": "[\"K:Economic costs and return\", \"K:Architecture and modularity\", \"K:Code reuse and duplication\"]",
  "editorial_chapters_json": "[\"C:C02\", \"C:C05\", \"C:C12\"]",
  "mapping_status": "Independent editorial mapping, not classification by the article authors or SE@PoliMi",
  "book_citation_status": "IN_PUBLISHED_NDSE2_BIBLIOGRAPHY",
  "node_class": "FOWLER_SERIES_ADDITIVE_OVERLAY_V12",
  "ndse2_actual_chapters_json": "[]",
  "ndse2_in_bibliography": true,
  "ndse2_citation_basis": "Final 13.7.1 active EN/ITA source tree; index links and nocite are not chapter citations",
  "historical_v12_book_citation_status": "NOT_ADDED_TO_BOOK"
}

• → F030 · The Economic Benefit of Refactoring
  F_BODY_LINKS_TO · edge f00332 · key MF_V12:F_BODY_LINKS_TO
• → W077 · Evaluating and Mitigating the Misguidance Effect of Buggy Code in LLM-Generated Unit Tests
  F_EDITORIAL_COMPARISON · edge f00397 · key MF_V12:F_EDITORIAL_COMPARISON
• → F006 · TDD with GitHub Copilot
  F_EDITORIAL_COMPARISON · edge f00398 · key MF_V12:F_EDITORIAL_COMPARISON
• → B091 · Quality Gates and Backpressure
  F_EDITORIAL_COMPARISON · edge f00399 · key MF_V12:F_EDITORIAL_COMPARISON

These retain the graph’s exact relation labels. A connection is not automatically a causal claim or an independent replication.

{
  "label": "TDD inside the agent loop - theater or actual value?",
  "kind": "practitioner_essay",
  "collection": "F",
  "authors": "Birgitta Böckeler",
  "authors_json": "[\"Birgitta Böckeler\"]",
  "publication_date": "2026-08-10",
  "date_provenance": "Dated entry on the supplied series index; article date also used where printed",
  "url": "https://martinfowler.com/articles/exploring-gen-ai/tdd-in-the-agent-loop.html",
  "publisher": "martinfowler.com",
  "series": "Exploring Generative AI",
  "series_url": "https://martinfowler.com/articles/exploring-gen-ai.html",
  "snapshot_date": "2026-10-13",
  "contribution": "Separates human-authored tests, a human review checkpoint and TDD entirely inside an agent loop. The small exploratory comparison found no clear advantage for fully internal TDD in model-judged quality or mutation scores.",
  "limitation": "Small greenfield tasks, imperfect adherence and an LLM-defined quality rubric. Reported aggregate tokens include cache rereads and are not equivalent to billed cost; no universal conclusion about human-led TDD follows.",
  "evidence_class": "Small exploratory agent evaluation",
  "source_locator": "Setup; results; test-first and red-green goals; costs; conclusions",
  "verification_scope": "Primary article byline and selected body passages inspected; no independent execution or statistical replication",
  "original_data_replicated": false,
  "new_manuscript_citation": false,
  "source_body_archived": false,
  "primary_stream": "T:Verification & review",
  "streams_json": "[\"T:Verification & review\", \"T:Economics\"]",
  "topics_json": "[\"K:Independent verification and oracles\", \"K:Evaluation methodology and baselines\", \"K:Economic costs and return\"]",
  "editorial_chapters_json": "[\"C:C03\", \"C:C07\", \"C:C12\", \"C:C13\"]",
  "mapping_status": "Independent editorial mapping, not classification by the article authors or SE@PoliMi",
  "book_citation_status": "IN_PUBLISHED_NDSE2_BIBLIOGRAPHY",
  "node_class": "FOWLER_SERIES_ADDITIVE_OVERLAY_V12",
  "ndse2_actual_chapters_json": "[]",
  "ndse2_in_bibliography": true,
  "ndse2_citation_basis": "Final 13.7.1 active EN/ITA source tree; index links and nocite are not chapter citations",
  "historical_v12_book_citation_status": "NOT_ADDED_TO_BOOK"
}

• → P070 · New trend: programming by kicking off parallel AI agents
  F_EDITORIAL_COMPARISON · edge f00400 · key MF_V12:F_EDITORIAL_COMPARISON
• → B095 · Multi-Agent Orchestration
  F_EDITORIAL_COMPARISON · edge f00401 · key MF_V12:F_EDITORIAL_COMPARISON

These retain the graph’s exact relation labels. A connection is not automatically a causal claim or an independent replication.

{
  "label": "An Accidental Blackboard",
  "kind": "practitioner_essay",
  "collection": "F",
  "authors": "Giles Edwards-Alexander",
  "authors_json": "[\"Giles Edwards-Alexander\"]",
  "publication_date": "2026-09-02",
  "date_provenance": "Dated entry on the supplied series index; article date also used where printed",
  "url": "https://martinfowler.com/articles/exploring-gen-ai/an-accidental-blackboard.html",
  "publisher": "martinfowler.com",
  "series": "Exploring Generative AI",
  "series_url": "https://martinfowler.com/articles/exploring-gen-ai.html",
  "snapshot_date": "2026-10-13",
  "contribution": "In a simulated airline exercise, shared plans and frequent repository updates became an improvised coordination surface for agents. The account connects this behavior with the blackboard pattern and describes resulting CI pressure.",
  "limitation": "A practice exercise, not a deployed airline system. The author explicitly questions reproducibility; the coordination mechanism was emergent rather than experimentally isolated.",
  "evidence_class": "Exploratory team exercise",
  "source_locator": "Emergent behavior; shared plans; repository as blackboard; CI trade-off",
  "verification_scope": "Primary article byline and selected body passages inspected; no independent execution or statistical replication",
  "original_data_replicated": false,
  "new_manuscript_citation": false,
  "source_body_archived": false,
  "primary_stream": "T:Engineering structure",
  "streams_json": "[\"T:Engineering structure\", \"T:System knowledge & memory\"]",
  "topics_json": "[\"K:Human-agent collaboration\", \"K:Agent context and persistent memory\", \"K:Long-running agent execution\"]",
  "editorial_chapters_json": "[\"C:C03\", \"C:C09\", \"C:C11\"]",
  "mapping_status": "Independent editorial mapping, not classification by the article authors or SE@PoliMi",
  "book_citation_status": "COMPANION_READING_NOT_IN_BOOK",
  "node_class": "FOWLER_SERIES_ADDITIVE_OVERLAY_V12",
  "ndse2_actual_chapters_json": "[]",
  "ndse2_in_bibliography": false,
  "ndse2_citation_basis": "Final 13.7.1 active EN/ITA source tree; index links and nocite are not chapter citations",
  "historical_v12_book_citation_status": "NOT_ADDED_TO_BOOK"
}

• → F030 · The Economic Benefit of Refactoring
  F_BODY_LINKS_TO · edge f00357 · key MF_V12:F_BODY_LINKS_TO
• → W095 · Beyond Code Generation: Reliability, Verification, and Cost Economics in the Agentic Software Development Lifecycle
  F_EDITORIAL_COMPARISON · edge f00402 · key MF_V12:F_EDITORIAL_COMPARISON
• → B106 · The Economics of Loop Engineering
  F_EDITORIAL_COMPARISON · edge f00403 · key MF_V12:F_EDITORIAL_COMPARISON
• → B109 · Choosing the Right Loop
  F_EDITORIAL_COMPARISON · edge f00404 · key MF_V12:F_EDITORIAL_COMPARISON
• ← NDSE2:C12 · Economics, Jobs, Risk and Evidence-Based Adoption
  NDSE2_CHAPTER_CITES · edge u20261018_01075 · key UPDATE_20261018:NDSE2_CHAPTER_CITES:EN
• ← NDSE2:C12 · Economics, Jobs, Risk and Evidence-Based Adoption
  NDSE2_CHAPTER_CITES · edge u20261018_01098 · key UPDATE_20261018:NDSE2_CHAPTER_CITES:ITA

These retain the graph’s exact relation labels. A connection is not automatically a causal claim or an independent replication.

{
  "label": "Are Harnesses Worth the Cost?",
  "kind": "practitioner_essay",
  "collection": "F",
  "authors": "Naresh Jain",
  "authors_json": "[\"Naresh Jain\"]",
  "publication_date": "2026-10-13",
  "date_provenance": "Dated entry on the supplied series index; article date also used where printed",
  "url": "https://martinfowler.com/articles/are-harnesses-worth-the-cost.html",
  "publisher": "martinfowler.com",
  "series": "Exploring Generative AI",
  "series_url": "https://martinfowler.com/articles/exploring-gen-ai.html",
  "snapshot_date": "2026-10-13",
  "contribution": "Compares development with and without a behavioral harness against a separate acceptance suite. On the selected service, the stronger model passed both conditions; a cost-optimized model passed more consistently with harness feedback, at greater per-attempt effort.",
  "limitation": "Ten paired runs per model on one unusually detailed greenfield Java brief. Guides and sensors changed together; estimated generation costs omit harness authoring, review and repairs. Final-gate-only use is proposed, not tested.",
  "evidence_class": "Small repeated-run implementation experiment",
  "source_locator": "Experiment setup; independent outer evaluation; model comparisons; economics; boundaries",
  "verification_scope": "Primary article byline and selected body passages inspected; no independent execution or statistical replication",
  "original_data_replicated": false,
  "new_manuscript_citation": true,
  "source_body_archived": false,
  "primary_stream": "T:Economics",
  "streams_json": "[\"T:Economics\", \"T:Verification & review\", \"T:Engineering structure\"]",
  "topics_json": "[\"K:Economic costs and return\", \"K:Independent verification and oracles\", \"K:Specifications and governance\"]",
  "editorial_chapters_json": "[\"C:C02\", \"C:C03\", \"C:C04\", \"C:C07\", \"C:C12\", \"C:C13\"]",
  "mapping_status": "Independent editorial mapping, not classification by the article authors or SE@PoliMi",
  "book_citation_status": "IN_PUBLISHED_NDSE2_BIBLIOGRAPHY",
  "node_class": "FOWLER_SERIES_ADDITIVE_OVERLAY_V12",
  "ndse2_actual_chapters_json": "[\"C:C12\"]",
  "ndse2_in_bibliography": true,
  "ndse2_citation_basis": "Final 13.7.1 active EN/ITA source tree; index links and nocite are not chapter citations",
  "historical_v12_book_citation_status": "NOT_ADDED_TO_BOOK",
  "historical_v12_new_manuscript_citation": false
}

• → F020 · Partner with the AI, throw away the code
  F_EDITORIAL_COMPARISON · edge u20261018_00030 · key UPDATE_20261018:F_EDITORIAL_COMPARISON

These retain the graph’s exact relation labels. A connection is not automatically a causal claim or an independent replication.

{
  "label": "Practitioner Voice",
  "kind": "practitioner_essay",
  "collection": "F",
  "authors": "Jim Highsmith",
  "authors_json": "[\"Jim Highsmith\"]",
  "publication_date": "2026-08-19",
  "updated_date": "2026-10-13",
  "latest_activity_date": "2026-10-13",
  "date_provenance": "Primary article Significant Revisions: original publication and later expansion distinguished",
  "url": "https://martinfowler.com/articles/practitioner-voice.html",
  "publisher": "martinfowler.com",
  "series": "",
  "series_url": "",
  "source_scope": "Adjacent martinfowler.com article; outside the original Exploring Generative AI series catalogue",
  "series_member": false,
  "snapshot_date": "2026-10-18",
  "contribution": "The October revision adds practical guidance for writing from experience and using AI while keeping the writer's judgment visible. Highsmith describes cognitive assistance, contrasts different authoring choices, and connects useful prose to lived consequences, audience, storytelling and iterative editing.",
  "limitation": "Personal experience and editorial guidance; no controlled evaluation of cognition, learning, writing quality or productivity. Do not treat the revision as a newly published October article.",
  "evidence_class": "Practitioner essay and personal experience",
  "source_locator": "The Craft of Practitioner Voice; How to use AI without losing your voice; Significant Revisions",
  "verification_scope": "Actual byline, original publication date and October 13 significant revision confirmed. No independent execution, causal evaluation or statistical replication.",
  "access_status": "FULL_PRIMARY_TEXT",
  "original_data_replicated": false,
  "new_manuscript_citation": false,
  "source_body_archived": false,
  "primary_stream": "T:Skills & learning",
  "streams_json": "[\"T:Skills & learning\",\"T:System knowledge & memory\"]",
  "topics_json": "[\"K:Human cognitive load\",\"K:Cognitive offloading and skill debt\",\"K:Ownership and knowledge transfer\"]",
  "editorial_chapters_json": "[\"C:C09\",\"C:C10\",\"C:C11\"]",
  "mapping_status": "Independent editorial mapping, not classification by the article authors or SE@PoliMi",
  "book_citation_status": "COMPANION_READING_NOT_IN_BOOK",
  "node_class": "FOWLER_SITE_EXTENSION_20261018",
  "addition_reason": "October 13 expansion on AI-assisted thinking and writing, absent from the series-only baseline. Scope expansion explicitly retained.",
  "atlas_added_on": "2026-10-18",
  "canonical_url": "https://martinfowler.com/articles/practitioner-voice.html",
  "online_companion_update": true,
  "ndse2_actual_chapters_json": "[]",
  "ndse2_in_bibliography": false,
  "ndse2_citation_basis": "Final 13.7.1 active EN/ITA source tree; index links and nocite are not chapter citations",
  "pre_reconciliation_book_citation_status": "NOT_ADDED_TO_BOOK",
  "manuscript_citation_status": "NOT_CITED_IN_PUBLISHED_NDSE2_13_7_1",
  "manuscript_citation_basis": "Absent from the final published NDSE2 13.7.1 active EN/ITA source tree and loaded bibliography"
}

• → F009 · Onboarding to a 'legacy' codebase with the help of AI
  F_EDITORIAL_COMPARISON · edge u20261018_00031 · key UPDATE_20261018:F_EDITORIAL_COMPARISON
• → F017 · Building Custom Tooling with LLMs
  F_EDITORIAL_COMPARISON · edge u20261018_00032 · key UPDATE_20261018:F_EDITORIAL_COMPARISON

These retain the graph’s exact relation labels. A connection is not automatically a causal claim or an independent replication.

{
  "label": "When Software Meets Hardware",
  "kind": "practitioner_essay",
  "collection": "F",
  "authors": "Kuldeep Singh",
  "authors_json": "[\"Kuldeep Singh\"]",
  "publication_date": "2026-10-15",
  "updated_date": null,
  "latest_activity_date": "2026-10-15",
  "date_provenance": "Primary article dateline and Significant Revisions",
  "url": "https://martinfowler.com/articles/software-meets-hardware.html",
  "publisher": "martinfowler.com",
  "series": "",
  "series_url": "",
  "source_scope": "Adjacent martinfowler.com article; outside the original Exploring Generative AI series catalogue",
  "series_member": false,
  "snapshot_date": "2026-10-18",
  "contribution": "An IoT modernization account connects physical test environments, device simulation, observability, recovery and continuous learning. The team used LLMs to explore unfamiliar areas, prototype experiments and develop engineering tools, including dashboards and migration trackers. The proposed benefit is cheaper experimentation and a shorter learning loop.",
  "limitation": "The article describes one team's practices and perceived benefits. It does not isolate AI's contribution or quantify productivity improvements. Physical-world validation remains part of the reported engineering process.",
  "evidence_class": "Qualitative engineering experience report",
  "source_locator": "Experimentation environment; physical testing; Use AI to shorten the learning loop; Capture discoveries; Significant Revisions",
  "verification_scope": "Byline, publication history and AI-specific section inspected. No independent execution, causal evaluation or statistical replication.",
  "access_status": "FULL_PRIMARY_TEXT",
  "original_data_replicated": false,
  "new_manuscript_citation": false,
  "source_body_archived": false,
  "primary_stream": "T:Engineering structure",
  "streams_json": "[\"T:Engineering structure\",\"T:Skills & learning\",\"T:Productivity\"]",
  "topics_json": "[\"K:Onboarding and developer fluency\",\"K:System understanding and rationale\",\"K:Developer workflow and adoption\",\"K:Independent verification and oracles\"]",
  "editorial_chapters_json": "[\"C:C02\",\"C:C05\",\"C:C09\",\"C:C10\",\"C:C13\"]",
  "mapping_status": "Independent editorial mapping, not classification by the article authors or SE@PoliMi",
  "book_citation_status": "COMPANION_READING_NOT_IN_BOOK",
  "node_class": "FOWLER_SITE_EXTENSION_20261018",
  "addition_reason": "New October 15 experience report with an explicit AI-assisted learning and engineering-tooling section, outside the original series.",
  "atlas_added_on": "2026-10-18",
  "canonical_url": "https://martinfowler.com/articles/software-meets-hardware.html",
  "online_companion_update": true,
  "ndse2_actual_chapters_json": "[]",
  "ndse2_in_bibliography": false,
  "ndse2_citation_basis": "Final 13.7.1 active EN/ITA source tree; index links and nocite are not chapter citations",
  "pre_reconciliation_book_citation_status": "NOT_ADDED_TO_BOOK",
  "manuscript_citation_status": "NOT_CITED_IN_PUBLISHED_NDSE2_13_7_1",
  "manuscript_citation_basis": "Absent from the final published NDSE2 13.7.1 active EN/ITA source tree and loaded bibliography"
}