# 论文骨架后续任务

> 本页由仓库源文件生成。请不要直接编辑本页；修改下方 source 文件后运行 `vp run docs:generate`。

> 中文支持说明：本页是中文站点镜像，保留论文标题、证据状态和来源字段的原始表述，避免未经核实的翻译改变证据边界。

> 来源: `paper/PAPER_SKELETON_NEXT_STEPS.md`

## Paper Skeleton Next Steps

Date: 2026-07-12

This file tracks manuscript-structure work originally derived from the 40-paper execution plan and updated after the SSVEP/MI additions. Current execution status is split between `reports/additional_pdf_processing_report.md` for corpus intake and `reports/docs_research_wiki_reorganization_plan.md` for docs/wiki execution. It follows the local paper-review workflow: paper cards and evidence matrix first, manuscript claims second.

## Current State

* The active corpus is now 73 local PDFs with 73 extracted text artifacts.
* All active paper cards are structured: `72 extracted` and `1 extracted from OCR`; no active extraction queue remains.
* `GRASP-006` was recovered via OCR on 2026-07-10 and should be used for classical hand-eye calibration rationale only.
* SSVEP and MI tracks have been expanded from classical foundations to recent AR/VR, transfer-learning, Riemannian, deep-learning, wearable, and rehabilitation trends.
* `library/EVIDENCE_MATRIX.md` now has one `verified` literature-formalism row and several `inferred` design-rationale rows.
* Related Work is draft-ready from the active 73-paper corpus.
* Method now has an interface skeleton: `system/runtime-interfaces.md` and `system/state-machine.md`.
* Experiments now have analysis and pre-data freeze skeletons.
* The docs site now exposes the SAH-BRI framework source page and the full Paper Kit skeleton.
* Results, user workload, closed-loop performance, low-channel performance, middleware binding, and project-specific hand-eye calibration claims remain blocked.

## Immediate Tasks

| Priority | Task | Target Files | Rule |
| --- | --- | --- | --- |
| P0 | Update manuscript status and writing boundaries | `paper/manuscript.md` | Done in this skeleton pass. |
| P0 | Build Introduction scaffold | `paper/sections/01-introduction.md` | Done; uses `verified` and `inferred` labels explicitly. |
| P0 | Build Related Work scaffold | `paper/sections/02-related-work.md` | Done; superseded by P0+P1 full prose draft. |
| P0 | Expand Method outline | `paper/sections/03-method.md` | Done; architecture, interfaces, state machine, and safety gates expanded. |
| P0 | Expand Experiment outline | `paper/sections/04-experiments.md` | Done; experiments and baselines expanded, no results claimed. |
| P1 | Extract remaining active cards | `library/paper_cards/*.md` | Done for all 73 active paper cards; see `library/PAPER_CARD_REVIEW_QUEUE.md`. |
| P1 | Recover calibration evidence | `library/paper_cards/GRASP-006.md` | Done via OCR; cite for classical rationale only. |
| P1 | Define project calibration protocol | `experiments/exp03-closed-loop-grasping/hand-eye-calibration.md` | Skeleton done; camera/robot/target/solver/station count/acceptance thresholds still need hardware binding. |
| P1 | Draft full Related Work prose | `paper/sections/02-related-work.md` | Done from the active 73-paper corpus; system-level claims remain marked as inference. |
| P1 | Expose full manuscript skeleton in docs | `vocs.config.ts`; `scripts/generate_vocs_docs.mjs` | Done; Paper Kit includes overview, next steps, and sections 1-7. |

## Section Readiness

| Section | Status | Next Action |
| --- | --- | --- |
| Introduction | `scaffold ready` | Convert paragraph plan into venue-style prose after target venue is chosen. |
| Related Work | `draft ready` | Review prose for target venue style and citation formatting. |
| Method | `interface skeleton expanded` | Bind runtime interfaces and state-machine records to actual middleware, hardware IDs, and timing values after specs are known. |
| Experiments | `analysis skeleton drafted` | Fill participant counts, trial counts, hardware identifiers, channel montage, ITR time definition, and numeric thresholds before data collection. |
| Results | `blocked placeholder` | Wait for experimental outputs. |
| Discussion | `limitation scaffold drafted` | Fill result interpretation after experiments and failure-mode logs exist. |
| Conclusion | `blocked placeholder` | Wait for Results/Discussion. |

## Current Missing Pieces

| Gap Class | Missing Piece | Blocks | Current File To Fill |
| --- | --- | --- | --- |
| Writing | target venue style and citation format | final Introduction / Related Work prose | `paper/sections/01-introduction.md`; `paper/sections/02-related-work.md` |
| Implementation binding | middleware choice, message serialization, clock synchronization, log storage | operational Method claims and reproducible trial logs | `experiments/pre-data-freeze-checklist.md`; `system/runtime-interfaces.md` |
| Hardware binding | EEG, display, camera/RGB-D, robot, gripper, workcell | hardware-specific timing, safety, and calibration claims | `experiments/hardware-inventory.md` |
| Calibration binding | target, solver, station count, transform convention, acceptance thresholds | physical robot grasping and hand-eye accuracy claims | `experiments/exp03-closed-loop-grasping/hand-eye-calibration.md`; `experiments/templates/threshold-freeze-record.md` |
| Experimental design | participant criteria, participant count, trial counts, randomization, training/test split | evidence-producing Exp1-Exp4 sessions | `experiments/pre-data-freeze-checklist.md`; `experiments/templates/session-record.md` |
| Analysis | statistical model, paired comparison strategy, multiple-comparison handling, final ITR `T` definitions | confirmatory results section | `experiments/analysis-plan.md` |
| Results | Exp1-Exp4 data, exclusions, confidence intervals, failure-mode counts | Results / Discussion / Conclusion | `paper/sections/05-results.md`; `paper/sections/06-discussion.md` |
| Docs maintenance | regenerated Vocs pages after source edits | current public documentation state | `vocs.config.ts`; `scripts/generate_vocs_docs.mjs`; `docs/pages/` |

## Claim Rules

* `verified`: can be stated as literature or project evidence.
* `inferred`: can be used as design rationale, not as final performance conclusion.
* `needs confirmation`: must remain a gap, protocol item, or future work item.

## Next Best Work Batch

1. Review Introduction and Related Work prose for target venue style and citation formatting.
2. Bind runtime interfaces, state-machine records, hand-eye calibration, and thresholds to actual hardware/middleware before writing implementation claims.
3. Fill `experiments/pre-data-freeze-checklist.md`, `experiments/hardware-inventory.md`, and threshold/session templates before data collection.
4. Collect Exp1-Exp4 data before drafting Results, result-dependent Discussion, or Conclusion claims.