Skip to content

Latest commit

 

History

History
463 lines (377 loc) · 11.2 KB

File metadata and controls

463 lines (377 loc) · 11.2 KB

ACM TOSEM Manuscript Submission Guide

Journal: ACM Transactions on Software Engineering and Methodology (TOSEM)
Manuscript ID: TOSEM-2026-0481
Status: Requires reformatting and resubmission
Type: Single-blind (authors identified, not anonymized)


✅ Checklist: What ACM Requires

Author Information

  • All authors listed on first page
  • Full names (not anonymized)
  • Author affiliations
  • Corresponding author contact info
  • ORCID numbers (optional but recommended)

Manuscript Structure

  • Title page with author info
  • Abstract (150-250 words)
  • Introduction
  • Related Work
  • Methodology
  • Results & Analysis
  • Discussion
  • Conclusions
  • Acknowledgments
  • References
  • Appendices (if needed)

Formatting

  • ACM template (Word or LaTeX)
  • Single-blind (authors visible)
  • Line numbers for review
  • Page numbers
  • Proper citation format
  • Figure/table captions

Submission Package

  • Main manuscript (PDF)
  • Supplementary materials (datasets, code)
  • Cover letter
  • Author declarations
  • Conflict of interest statement

📝 Manuscript Structure for SigMap Paper

Title Page

Title: Large-Scale Analysis of AI Context Extraction Across 240 Open-Source Repositories

Authors:
- Manoj Mallick, [Affiliation]
  Email: [email]
  ORCID: [if available]

Corresponding Author:
- Manoj Mallick
  Email: [email]
  Phone: [phone]

Acknowledgments:
[List collaborators, funding sources, etc.]

Abstract (200-250 words)

Focus:
- Problem: Token limits in LLM context management
- Solution: SigMap - systematic evaluation of context extraction
- Dataset: 405 open-source repositories, 30+ languages
- Key findings: 96.2% token reduction, language patterns matter
- Contribution: Empirical foundation for context extraction research
- Data availability: CC-BY-4.0 licensed, publicly available

Keywords: Context extraction, LLM optimization, Code compression, 
Empirical software engineering, AI-assisted development

1. Introduction (1000-1200 words)

Sections:
1.1 The Problem
    - LLM token limits
    - Context management challenge
    - Industry relevance

1.2 Motivation
    - Why context extraction matters
    - Gap in empirical knowledge
    - Opportunity for systematic study

1.3 Our Approach
    - Large-scale evaluation
    - 405 diverse repositories
    - Reproducible methodology

1.4 Key Contributions
    - Largest empirical study of context extraction
    - Language-specific patterns identified
    - Monorepo optimization insights
    - Complete reproducibility package

1.5 Paper Organization

2. Related Work (800-1000 words)

Sections:
2.1 Context Management in LLMs
    - Token limits and their impact
    - Existing approaches
    - Limitations of current methods

2.2 Code Compression & Summarization
    - Prior work on code summarization
    - Token reduction techniques
    - Language-specific approaches

2.3 Empirical Software Engineering
    - Large-scale code analysis studies
    - Multi-language evaluation
    - Reproducibility in SE research

2.4 AI-Assisted Software Engineering
    - LLM integration in development
    - Context optimization
    - Tool effectiveness

2.5 Gap Analysis
    - What's missing from prior work
    - Why systematic study needed
    - Our contribution

3. Methodology (1200-1500 words)

Sections:
3.1 Research Questions
    - RQ1: How effective is context extraction across languages?
    - RQ2: What factors affect compression rates?
    - RQ3: Can findings generalize?
    - RQ4: What's the practical impact?

3.2 Dataset & Repository Selection
    - Stratified sampling approach
    - 405 repositories selected
    - Language distribution
    - Size distribution
    - Diversity metrics

3.3 Execution Environment
    - Hardware specs (c2-standard-8)
    - Software stack
    - SigMap tool version
    - Network conditions

3.4 Benchmark Modes
    - Health check
    - Baseline measurement
    - Standard analysis
    - Comprehensive reporting
    - Deep analysis
    - Mode parameters

3.5 Metrics & Measurement
    - Token reduction percentage
    - Compression ratio
    - Execution time
    - Success rate
    - Data completeness

3.6 Quality Assurance
    - Validation procedures
    - Error handling
    - Data integrity checks
    - Reproducibility verification

4. Results (1500-2000 words)

Sections:
4.1 Overall Statistics
    - 405 repositories analyzed
    - 2,025+ operations executed
    - 96.2% average token reduction
    - Success rate: 99.6%
    - Data completeness: 100%

4.2 Language-Specific Analysis
    - Python results (96.2% ± 1.8%)
    - JavaScript results (92.1% ± 4.2%)
    - Go, Rust, Java results
    - Comparative analysis
    - Statistical significance

4.3 Repository Size Analysis
    - Token reduction by size
    - Size doesn't predict compression
    - Other factors more important

4.4 Monorepo Analysis
    - 45 monorepos identified
    - Specialized handling impact
    - 2-3% improvement potential
    - Module count analysis

4.5 Framework Impact
    - Framework-specific patterns
    - React vs Vue (JavaScript)
    - Django vs FastAPI (Python)
    - Design pattern effects

4.6 Methodology Validation
    - Published (240) vs Extended (405) comparison
    - Identical 96.2% average
    - Findings generalize
    - Methodology robustness

5. Discussion (800-1200 words)

Sections:
5.1 Interpretation of Findings
    - Language organization matters
    - Code idioms primary factor
    - Framework diversity effects
    - Size is not determinant

5.2 Implications for Researchers
    - Baseline for future work
    - Language-specific optimization
    - Research directions

5.3 Implications for Tool Builders
    - Realistic token reduction targets
    - Monorepo optimization opportunity
    - Language-specific strategies

5.4 Limitations
    - Shallow clones (depth=1)
    - Single execution run
    - Tool-specific results
    - Language detection limitations
    - 300-second timeout

5.5 Threats to Validity
    - Selection: diverse sampling
    - Measurement: documented metrics
    - Construct: multiple benchmark modes
    - Internal: controlled environment
    - External: 405 repos, 30+ languages

5.6 Generalizability
    - Findings hold across sample sizes
    - Multiple languages represented
    - Open-source + production repos
    - Different domains covered

6. Conclusions (300-500 words)

- Summary of contributions
- Key findings recap
- Implications
- Future work opportunities
- Impact on field

7. Acknowledgments

Funding sources
Collaborators
Tool authors (SigMap team)
Infrastructure providers
Community contributors

8. References (30-50 citations)

Format: ACM style

Key papers to cite:
- LLM context management
- Code summarization
- Software engineering empirical studies
- Reproducible research
- AI-assisted development

9. Supplementary Materials (Optional)

- Complete dataset (CSV, JSON, JSONL, SQL)
- Reproducibility scripts
- Hardware configuration details
- Language statistics
- Benchmark output samples
- Extended tables/charts

📋 ACM Formatting Requirements

Manuscript Format

  • Word: Use ACM Master Article Template
  • LaTeX: Use acmart.cls class
  • PDF: Single-column format
  • Font: 10pt minimum
  • Spacing: Double-spaced for review

Author Information

First Author Name, Affiliation
Second Author Name, Affiliation
Corresponding Author: Name, Email, Phone

Headers & Footers

  • Page numbers: bottom center
  • Running head: your short title
  • Line numbers: for review version

Figures & Tables

  • High resolution (300+ DPI)
  • Captions with descriptions
  • References in text
  • ACM-approved colors
  • Legible fonts

Citations

  • ACM reference style
  • Include DOIs when available
  • Consistent formatting
  • Complete author lists

🔗 References & Links

ACM TOSEM:

ACM Style:

Supplementary:


📝 Cover Letter Template

Dear Editor-in-Chief,

We are submitting a revised manuscript titled "Large-Scale Analysis of AI 
Context Extraction Across 240 Open-Source Repositories" for publication in 
ACM Transactions on Software Engineering and Methodology.

[Brief description of contribution]

This work:
- Provides empirical foundation for context extraction research
- Includes complete reproducible methodology
- Releases public dataset (CC-BY-4.0)
- Identifies language-specific optimization opportunities
- Validates findings across 240→405 repositories

All required materials are included:
✓ Reformatted manuscript (ACM template)
✓ Author information (non-anonymized)
✓ Complete datasets (CSV, JSON, JSONL, SQL)
✓ Reproducibility scripts
✓ Research papers (primary + extended)

We have no conflicts of interest to declare.

The authors are: [List all authors with affiliations]
Corresponding author: [Name, email, phone]

Thank you for considering this submission.

Sincerely,
Manoj Mallick
[Affiliation]

✅ Before Resubmitting: Checklist

Formatting

  • Used ACM template (not your own format)
  • Non-anonymized (authors visible)
  • Line numbers present
  • Page numbers present
  • 10pt+ font, readable
  • All figures high resolution
  • References in ACM format

Content

  • Abstract: 150-250 words
  • Introduction: ~1000-1200 words
  • Related work: ~800-1000 words
  • Methodology: ~1200-1500 words
  • Results: ~1500-2000 words
  • Discussion: ~800-1200 words
  • Conclusions: ~300-500 words
  • References: 30-50 citations

Submission Package

  • Manuscript PDF
  • Cover letter
  • Author declarations
  • Conflict of interest statement
  • Supplementary materials
  • Data availability statement

Final Quality Check

  • Spell check complete
  • Grammar reviewed
  • Citations verified
  • Figures tested for clarity
  • Tables readable
  • All cross-references correct
  • No anonymization present

🚀 Submission Steps

  1. Download ACM template from https://www.acm.org/publications/submission-templates
  2. Reformat manuscript using template
  3. Add author information (non-anonymized)
  4. Prepare supplementary materials (datasets, scripts)
  5. Write cover letter (see template above)
  6. Final review using checklist
  7. Submit via TOSEM submission system
  8. Confirm receipt from editorial office

Expected review timeline: 3-4 months


📞 Support Resources


Ready to reformat and resubmit? Use this guide to ensure compliance with ACM TOSEM requirements.