How hmntsharma/draft-ietf-grow-bmp-tcp-ao works
Project type and purpose
• This repo is *not* a traditional full-stack application but an IETF Internet-Draft authoring and publishing workflow.
• Purpose: write and maintain the “TCP-AO Protection for BGP Monitoring Protocol (BMP)” draft, build HTML/plain-text outputs, and automate publication (GitHub Pages, Datatracker, archives).
File structure
• Top-level content:
• draft-ietf-grow-bmp-tcp-ao.md (source of the specification)
• Makefile (build commands)
• .github/workflows/... (CI/CD manifests)
• Supporting metadata: README.md, CONTRIBUTING.md, LICENSE.md
• No traditional “src/backend/frontend,” but there is a clear “source → build → publish” pipeline.
• Key patterns:
• A markdown source transformed by standard I-D toolchain (per SETUP.md)
• CI workflows for continuous publishing, archiving, license checks, and GitHub Pages deployment
README for additional insights
• Confirms build via make
• References external toolchain (I-D template)
• Points at GitHub Pages and IETF Datatracker integrations
Main components
a. Authoring Environment
• Authors write/edit draft-*.md locally or via GitHub Web Editor
• Local toolchain: Pandoc/iddraft scripts invoked through make
b. Version Control & CI
• GitHub Repo hosts source and workflows
• Workflows: publish.yml (build & deploy), ghpages.yml (push to Pages), archive.yml, update.yml, LICENSE.md check
c. Build System
• Makefile orchestrates conversion of .md → .txt/.html
d. Publication Targets
• GitHub Pages (editor’s copy)
• IETF Datatracker (individual draft)
• Archive repositories
e. External Dependencies
• I-D Template toolchain (martinthomson/i-d-template)
• Pandoc or other formatting tools
Relationships & interactions
• Author commits → pushes to GitHub → triggers CI workflows
• CI invokes make → generates HTML/Text artifacts
• Artifacts deployed to GitHub Pages and/or synced to IETF Datatracker via automated scripts
• LICENSE check workflow validates license file on each push
• Archive workflow captures snapshots
Architectural patterns & design principles
• GitOps-style automation: all publication steps defined as code in .github/workflows
• Separation of concerns: content (markdown) vs. build logic (Makefile) vs. deployment logic (CI workflows)
• Infrastructure as code: CI manifests declare deployment targets and triggers
Detail level
• Do not over-complicate: capture each unique logical component
• How content changes propagate end-to-end
• Conditional steps (e.g., license checks, archive triggers)