Boost.JSON Competitive Market Research

Overview

This repository contains all artifacts for the Boost.JSON Competitive Market Research project. The goal is to produce a 10–15 page analysis comparing Boost.JSON against three other widely used C++ JSON libraries:

• Glaze
• simdjson
• JSON for Modern C++ (nlohmann/json)

Completed Study (start here)

The primary entry point for this repository is the completed technical overview:

• 05_drafts/Boost_JSON_Technical_Overview.md

Structure

01_charter/           Project charter and scope documents.
02_sources/           Evidence log (CSV + MD).
03_feature_matrix/    Feature matrix (CSV + MD).
04_benchmarks/        Benchmark plans and results.
05_drafts/            Draft report sections.
90_prompts/           Agent prompt templates.
99_admin/             Repo metadata and contributor guidelines.

Workflow Summary

1. Evidence-first — All claims originate from 02_sources/EvidenceLog.csv.
2. Matrix-driven writing — Narrative content is generated from 03_feature_matrix/FeatureMatrix.csv.
3. Red-team review — Drafts are checked for factual grounding before approval.
4. Versioning — GitHub used as the single source of truth; Notion hosts stakeholder snapshots.

How to Use

• To contribute evidence: append new rows to 02_sources/EvidenceLog.csv following 02_sources/EvidenceLog.md rules.
• To update the matrix: only after evidence has been reviewed and approved.
• To generate narrative sections: use 90_prompts/agent_writer.md with the relevant matrix and evidence subsets.
• To review: use 90_prompts/agent_redteam.md to fact-check drafts.

Quickstart — Hybrid AutoSlice (Deep Research + Automated)

1. Create a new chat and paste the kickoff from 90_prompts/agent_autoslice_hybrid.md.
2. When prompted, paste up to four sources using the provided template (one per library if possible).
3. The agent will validate, then return four artifacts: evidence_csv, matrix_row_csv, draft_md, redteam_md.
4. Apply outputs per 07_slices/parse_speed__autoslice_poc.runbook.md (auto_commit=false by default).

Public scope and draft visibility

This repository is now public. In-progress drafts are provided for transparency and community review. Drafts may change and should not be cited without confirming Evidence IDs in 02_sources/EvidenceLog.csv.

Drafts in progress

• Executive summary (draft): 05_drafts/01_executive_summary.md
• Feature matrix narrative (draft): 05_drafts/02_feature_matrix_narrative.md
• Performance — Parse Speed (draft): 05_drafts/parse_speed.md
• Streaming Support (draft): 05_drafts/streaming_support.md
• Red-team checklists (drafts): 05_drafts/parse_speed_redteam.md, 05_drafts/streaming_support_redteam.md, 05_drafts/02_feature_matrix_narrative%20redteam.md

For the automation/runbook prompts used to generate these drafts, see 90_prompts/ and 07_slices/.

Licensing

Unless otherwise noted, content is shared under the repository license. This repository is the deliverable of the study; no PDF production is planned. Do not treat drafts as finalized deliverables; verify all claims against 02_sources/EvidenceLog.csv and 03_feature_matrix/FeatureMatrix.csv.