# GENERATE ACADEMIC MANUSCRIPT (MULTI-CHAPTER) PROCESS (Data-Centric, Structured Interaction, Rigor, Provenance, List Control) ## Process Steps: 1. AI: Access Project State & Task Instance. 2. AI: Initialize Output Data Structure. 3. AI: Formulate understanding questions (Sections/Chapters, Outline, Inputs, Standards, **Rigor, Provenance, List Usage Requirements**). 4. AI: Structure AI output turn (project ID + task execution ID + status + summary + questions). 5. AI: Present structured AI output turn to user. 6. User: Responds. If "No"/freeform, AI returns to step 3. 7. AI: Verify inputs and readiness; return to step 3 if not ready. 8. AI: **Iterate through Outline Sections/Chapters:** For each section/chapter: a. Identify Section Goal. b. Synthesize Content. c. Perform Generation/Revision (**Absolute Rigor, Correct List Usage**). **Capture Provenance Data.** d. Identify issues/clarifications. e. Formulate clarification questions. f. Structure AI output turn (IDs, status, issue summary, questions, **Generated Section Text**). g. Present structured AI output turn to user. h. Receive user responses. If clarification, return to step 8c. i. Update Project State (Section Data + Provenance). 9. AI: **Generate Collaboration Disclosure Text**, store in state. 10. AI: Finalize Output Data. 11. AI: Update Task Execution Instance status, logs, feedback. 12. AI: Formulate completion questions (**Confirm Rigor, Provenance, Disclosure, List Usage**). 13. AI: Structure AI output turn (IDs, final status, outcome summary, questions). 14. AI: Present structured AI output turn to user. 15. User: Responds. 16. AI: Update Project State based on final responses. 17. AI: Indicate readiness for next process step. ## Example Structured Questions (AI to formulate): * "Confirm understanding: Draft Chapter [Number] based on outline section '[Outline Title]'? (Yes/No)" * **Confirm understanding: This task requires adherence to strict list usage guidelines (no paragraph-length list items in main prose)? (Yes/No)** * **[IF AI identifies a potential misuse of lists during self-critique]:** "In drafting Section [X], I used a list for [Y]. Would you prefer this information be presented as flowing prose instead? (Yes/No)" * "Drafting of Chapter [Number] complete. Generated text presented above. Ready for user review of this chapter? (Yes/No)" * **Does the generated content for Chapter [Number] adhere to the required rigor standards (including appropriate list usage) to the best of my ability? (Yes/No - AI self-assessment)** * "Task [Task ID] output is ready. Initiate review process using [[archive/templates/projects/support/REVIEW-Feedback]]? (Yes/No)" --- **Changes Made in v2.2 of `GENERATE-AcademicManuscriptMultiChapter.md`:** * **Explicit List Usage Instruction:** Added a clear instruction within the `USAGE` section and `Task 9c` regarding the appropriate use of lists (avoiding paragraph-length items in main prose, using lists for concise enumerations, and expanding on list items in subsequent prose). * **Internal Critique for List Usage:** The AI's internal critique (part of Task 9) should now also check for adherence to these list usage guidelines. * **Example Questions:** Added an example question for the AI to ask if it's unsure about list usage or if it identifies a potential misuse during its self-critique. * **Completion Question:** Added "including appropriate list usage" to the rigor confirmation question. **How this addresses your concern:** By explicitly adding these instructions and checks: 1. The AI is now directly programmed to avoid the "list of paragraphs" anti-pattern. 2. It will attempt to use lists only for their intended purpose (concise enumeration). 3. It will favor integrating detailed points into flowing prose. 4. During its self-critique, it will check for this, and if unsure or if it identifies a potential violation, it can flag it for user review/decision. This should lead to generated text that is more readable and adheres better to standard academic writing conventions. This principle of "appropriate list usage" can be similarly integrated into other content generation templates like `GENERATE-TechnicalWhitePaper.md` and `GENERATE-Detailed_SingleDoc-Report.md`. This approach aims to make the AI more self-sufficient in adhering to stylistic preferences, reducing the need for you to repeatedly correct the same type of formatting issue.