Academic Research Writing

Overview

Comprehensive toolkit for writing and reviewing computer science research papers. Combines paper writing workflows, manuscript review processes, clarity principles, and formatting standards.

When to Use

Writing Mode:

• Writing new research papers (conference, journal, thesis)
• Creating survey/review papers
• Structuring technical contributions

Review Mode:

• Reviewing/editing existing manuscripts
• Pre-submission polish
• Addressing reviewer comments
• Collaborative editing

Both Modes:

• Improving academic writing clarity
• Preparing IEEE/ACM submissions
• Learning academic writing conventions

Mode Selection

User request received
        |
        v
Is this about WRITING new content or REVIEWING existing content?
        |
        +---> Writing new paper -----> Use Writing Workflow
        |                              (references/writing-workflow.md)
        |
        +---> Reviewing/editing -----> Use Review Workflow
        |     existing manuscript      (references/review-workflow.md)
        |
        +---> Both/unclear ----------> Start with Review Workflow
                                       to assess, then write

Quick Reference

Writing a Paper

1. Clarify scope - topic, venue, format (IEEE/ACM)
2. Create outline - section-by-section plan
3. Draft core sections - methodology first, then results
4. Write supporting sections - intro, related work, discussion
5. Add citations - 15-20+ references
6. Review & polish - use checklists

See: references/writing-workflow.md

Reviewing a Manuscript

1. Extract core message - one sentence summary
2. Structural pass - overall organization
3. Section reviews - intro, results, discussion
4. Scientific clarity - claims, evidence, hedging
5. Language polish - terminology, voice
6. Formatting check - journal compliance

See: references/review-workflow.md

Core Resources

| Resource | Purpose | |----------|---------| | references/writing-workflow.md | 6-step paper writing process | | references/review-workflow.md | 8-step manuscript review process | | references/narrative-framework.md | Section-by-section narrative structure (Problem->Solution->Evidence) | | references/clarity-principles.md | Gopen & Swan sentence-level clarity | | references/academic-phrasebank.md | Common academic phrases by section | | references/cs-conventions.md | CS-specific writing conventions | | references/section-checklists.md | Combined quality checklists | | references/ieee-formatting.md | IEEE formatting specifications | | references/acm-formatting.md | ACM formatting specifications |

Templates

| Template | Purpose | |----------|---------| | templates/paper-structure.md | Introduction arc, results paragraph, discussion templates | | templates/methodology.md | Core message extraction, structural assessment, language guidelines |

Evaluation

Use evaluators/rubric.json for quality scoring:

• Structure and Organization (weight: 1.0)
• Scientific Rigor (weight: 1.2)
• Language and Clarity (weight: 1.0)
• Section-Specific Quality (weight: 1.0)
• Formatting Compliance (weight: 0.8)
• Citation Quality (weight: 0.8)

Minimum threshold: Average score >= 3.5

Seven Core Principles

1. Clarity over cleverness - Scientific clarity beats stylistic elegance
2. Narrative shapes comprehension - Structure determines understanding
3. Audience dictates tone - Expert vs. general requires different framing
4. Format signals credibility - Professional formatting reflects rigor
5. Claims require evidence - Strong assertions need strong data
6. Each section has a job - Intro sells, Results show, Discussion interprets
7. Constraints shape structure - Word limits determine emphasis

Guardrails

Critical requirements:

1. Preserve author voice - edit for clarity, don't rewrite
2. Claims match data - flag overclaiming immediately
3. Quantitative rigor - statistics for all comparisons
4. Logical flow - clear transitions between sections
5. Appropriate hedging - match evidence strength
6. Consistent terminology - same term for same concept

Common pitfalls to avoid:

• Overclaiming ("proves" when data only suggests)
• Missing context (results without interpretation)
• Buried lede (important findings hidden)
• Inconsistent terms (alternating synonyms)
• Vague descriptions ("some increase" vs "3-fold increase")

External Guides

| Guide | Purpose | |-------|---------| | external-guides/how-to-write-a-paper.md | Practical conference paper structuring guidelines (Introduction paragraphs, experiments, tables/figures) |

PDF Templates

• assets/full_paper_template.pdf - IEEE template
• assets/interim-layout.pdf - ACM template