📖 Keshav Three-Pass Reading Log

Keshav’s Three-Pass Method (S. Keshav, University of Waterloo, 2007):

• Pass 1 (5–10 min): bird’s-eye view using title, abstract, intro, headings, conclusions
• Pass 2 (≤1h): grasp content, note figures, mark unread references
• Pass 3 (4–5h): re-implement / virtually reconstruct the paper

[!info] Overview

Used in MIT AI Lab Working Paper 316 (“How to Do Research”) as reading guidance for graduate students. Integrated here as the PUMA paper-reading framework.

---

Reading Queue (Pass 1 pending)

TABLE title AS "Title", first-author AS "Author", year AS "Year", relevance AS "⭐"
FROM #literature
WHERE status = "to-read"
SORT relevance DESC

---

In Progress (Pass 2)

TABLE title AS "Title", first-author AS "Author", year AS "Year"
FROM #literature
WHERE status = "reading"

---

Completed — Ready for Permanent Notes (Pass 3 done)

TABLE title AS "Title", first-author AS "Author", year AS "Year", topic AS "Topic"
FROM #literature
WHERE status = "reviewed"
SORT year DESC

---

All Literature

TABLE title AS "Title", first-author AS "Author", year AS "Year", status AS "Status", relevance AS "⭐", topic AS "Topic"
FROM #literature
SORT relevance DESC, year DESC

---

Manual Log (for papers not yet in Dataview)

Paper	Author	Year	Pass 1	Pass 2	Pass 3	Permanent Note
Reproducibility of LLM Studies in SE	Angermeir	2025	✅	✅	✅	PN-KeyConcepts-Agents-Reproducibility-RedTeam
CoGEE: Story Point Estimation	Tawosi	2024	✅	✅	🔄	PN-IssueTriage-StoryPoints
How Big Things Get Done	Flyvbjerg	2023	✅	✅	✅	PN-IssueTriage-StoryPoints
Chain-of-Thought Prompting	Wei	2022	✅	✅	✅	PN-CoT-FewShot-Prompting
TAWOS Dataset	Tawosi	2022	✅	✅	✅	LN-Datasets-JiraSR-TAWOS
Design Science in IS Research	Hevner	2004	✅	✅	✅	PN-DSR-SLR-Methods
Cognitive Agents for Agile PM	Spichkova	2025	✅	🔄	⬜	—
Jira SR Dataset	Ortu	2015	✅	✅	✅	LN-Datasets-JiraSR-TAWOS
PM-LLM-Benchmark	Berti	2024	✅	🔄	⬜	—
AI in Project Management 2019–2024	Manzoor	2025	✅	✅	🔄	—
Request Formats and Effort Estimation	Calikli	2025	✅	✅	✅	PN-CoT-FewShot-Prompting
Energy and Policy for Deep Learning	Strubell	2019	✅	✅	✅	Carbon-Tracking-Log
Experimentation in SE	Wohlin	2012	✅	✅	✅	PN-Wilcoxon-FINER-Cornell-PRISMA
Local LLMs for Sprint Estimation	Yonathan	2025	✅	🔄	⬜	—
Language Models are Few-Shot Learners	Brown	2020	✅	✅	✅	PN-CoT-FewShot-Prompting

Legend: ✅ Done · 🔄 In Progress · ⬜ Not started

---

Keshav Pass-1 Quick Template

For fast capture during Pass 1, use:

**Paper**: [Title]
**Author**: [Last name, Year]
**Category**: measurement | analysis | description | proposal
**Context**: Related to [papers/theories]
**Correctness**: Assumptions [seem valid | questionable because...]
**Contributions**: 1) ... 2) ... 3) ...
**Clarity**: [clear | confusing | jargon-heavy]
**Decision**: Read Pass 2? [YES / NO — reason]

---

Permanent Notes Generated from Reading

LIST
FROM "30 - Permanent"
WHERE type = "permanent"
SORT file.ctime DESC
LIMIT 20