Fork me on GitHub

Technical Strategy: Debrief Rewrite Project

1. Vision and Scope

The Debrief rewrite project aims to modernise the existing Java-based desktop application into a modular, browser-first platform. The new architecture will improve:

  • Maintainability and developer onboarding
  • Integration with MOD digital infrastructure (cloud, identity, security)
  • Support for collaboration and distributed workflows
  • Reproducibility through pipeline capture (RAP)
  • Extensibility through AI plugins and modular tools

The system is designed to operate in multiple deployment contexts:

  • Standalone: Minimal install with local data access
  • Collaborative: Shared services via MODNET or site-local servers
  • AI-enabled: Integration of automated insights and agent-driven workflows

2. Architectural Overview

Component Map

  • Core Sub-Systems (required for minimal standalone operation)
  • Shared Services (optional, scalable extensions)
  • AI Supervision (optional automation layer)

Runtime Environments

  • Browser-based UI (React)
  • Electron-based shell (for deployed/offline use)
  • Static STAC server (local)
  • REST-based services (networked mode)

Storage Models

  • Local FeatureCollection files (GeoJSON + audit)
  • SQLite (optional for RAP, participant indexing)
  • STAC-based server for central access to spatial/temporal data

3. Sub-System Summary

Core Sub-Systems

  1. Client UI
  2. STAC Server (Static / File-Based)
  3. Import Service
  4. Export Service
  5. Pipeline Engine

Shared Services

  1. Platform Library
  2. Pipeline Processor
  3. STAC Server (Dynamic / Server-Based)

Collaborative Services

  1. Authentication
  2. Commenting
  3. Presence Locking
  4. Analysis Dashboard
  5. Wargame Metadata

AI Supervision

  1. LLM Supervisor
  2. MCP Agent Registry
architecture-beta group core(server)[Core] service db(database)[Client UI] in core service disk1(disk)[Data store] in core service disk2(internet)[Import service] in core service server(internet)[Export service] in core service pipes(server)[Pipeline Engine] in core group subs(Internet)[Subsystems] group llm(internet)[LLM] in subs service l1(cloud)[LLM Supervisor] in llm service l2(cloud)[MCP Agent Registry] in llm group shared(internet)[Shared] in subs service d8(database)[Platform Library] in shared service d11(server)[Pipeline Processor] in shared service d12(database)[Shared Data Store] in shared group collab(internet)[Collaborative] in subs service d13(server)[Authentication] in collab service d6(server)[Commenting] in collab service d7(server)[Presence Locking] in collab service d9(cloud)[Analysis Dashboard] in collab service d10(database)[Wargame Metadata] in collab

Each sub-system may be deployed independently based on operational needs.

4. Integration Patterns

  • REST APIs between services
  • File-based inputs/outputs (.rep, .dpf, .geojson)
  • STAC standard for discoverability and asset metadata
  • RAP pipelines stored inline or externally
  • FeatureCollections reference wargames and serials via wargameId, serialId

5. Security and Identity

  • Standalone installs run with OS-level file permissions
  • Shared deployments integrate with OIDC or PKI identity providers
  • Presence and locking managed at the FeatureCollection level
  • Annotations may be public or private; private annotations may be stored locally or in a presence service

6. Collaboration Modes

  • Standalone: Users load local files, no presence/locking, private annotations only

  • Team-based wargame analysis:

    • Wargame/serial metadata structures loaded
    • FeatureCollections tagged by serial
    • Dashboard shows status, comments, and edit ownership

  • Review workflows:

    • View-only dashboards for reviewers
    • Unresolved comments, change logs, audit trails visible

7. Extensibility Strategy

  • MCP agents are schema-described REST plugins
  • UI extensions provided via manifests + RJSF forms
  • RAP steps track tool execution, parameters, and results
  • Supervisor orchestrates agents based on analyst tasks

8. Export and Review Outputs

  • Storyboard Export:

    • Timeline-based, with viewports and visible annotations
    • Events table per step
    • Export as HTML bundle, PPT, screenshots, or MP4

  • Other Formats:

    • GeoJSON, KML, CSV
    • Static viewer with timeline controls
    • RAP summaries (HTML table)

9. Roadmap and Deployment Strategy (Optional)

  • Phase 1: NATO prototype (backend foundation + retro-style UI)
  • Phase 2: MOD adoption (minimal core system)
  • Phase 3: Shared services and dashboard integration
  • Phase 4: AI supervision and RAP automation

Appendices

  • Mermaid diagrams (Architecture, RAP, Dashboard flows)
  • Wargame/Serial JSON Schema (in future)
  • RAP step format and pipeline export model