TODO

  • Audit which ProsperPath cases can support a defensible 15-30 item decision-situation set
  • Tighten the pilot scoring rubric so context completeness, traceability, missing-information awareness, and decision usefulness are less overlapping
  • Define the exact inclusion and exclusion rules for the next experiment round before writing any results section
  • Finalize problem definition
  • Refine target contribution framing
  • Collect 15-20 core references
  • Write reading notes for priority papers
  • Define graph schema
  • Define signal categories
  • Define evaluation setup
  • Expand the source base on SME owner-manager decision pain points beyond AI-adoption reviews
  • Clarify whether the target concept should be framed as scenario memory, episodic business memory, workflow memory, case memory, or decision traceability
  • Find stronger enterprise or BPM literature on context reconstruction, event logs, and decision traceability for business decisions
  • Collect stronger empirical studies on informal SME decision processes, timely information access, and fragmented decision inputs
  • Validate the closest established terms for scenario-level business context, especially episodic memory, workflow memory, case-based reasoning, and decision traceability
  • Choose the narrowest defensible positioning: readiness gap, traceability gap, or episode-level context reconstruction gap
  • Refine the review-analysis article into a citation-tight related-work narrative with a final source audit
  • Draft introduction
  • Draft related work
  • Draft method skeleton
  • Convert the Method section into a citation-backed architecture narrative by mapping each design block to either implementation evidence (prosperpath) or formal literature support
  • Compress the Section 4 evaluation-design notes into a paper-ready experiments section after the next 15-30 case round is defined
  • Turn experiments/decision_episode_targeted_retest_results.md into a narrow paper subsection on targeted stress-test superiority versus combined_rrf
  • Decide whether the full regression table belongs in the main text or appendix, since it does not show universal DecisionEpisode superiority
  • Decide how manuscript/05_experiments/01-03 should map into the final Section 5 assembly and whether the full regression table should move to an appendix-style note
  • Add a citation-tight pass and export-ready assembly for manuscript/sections/short_paper_conference.md
  • Add a second-rater or structured follow-up validation plan for the short-paper scoring rubric, especially for over-filtering and omitted-constraint cases