We read every piece of feedback, and take your input very seriously.
To see all available qualifiers, see our documentation.
update wiki for vectorized engine + JS round-trip batching (April 2026) - Home: update description, date, add architecture section - Async-Dashboard-Approaches: add Approach 5 (Vectorized Batched WebR), update comparison table with speed/UI columns, fix decision flowchart - How-Async-Dashboard-Uses-Crew: fix dead links, add notes that browser dashboard no longer uses run_simulation() or crew, clarify scope Staleness check after session with 6 commits touching dashboard, CI, pkgdown config. See #199 for the staleness check rule. Co-Authored-By: Claude Opus 4.6 (1M context) <noreply@anthropic.com>
fix: Comprehensive wiki consistency review Async-Dashboard-Approaches.md: - Remove incorrect "~10-20 messages" claims (actual: O(N)) - Add sync mode (workers=0) as WebR-compatible option - Add chunked mode as RECOMMENDED approach - Mark dynamic/mirai_dynamic as DEPRECATED with explanation - Fix WebR compatibility claims (workers=0 IS compatible) - Restructure to show all 4 modes clearly How-Async-Dashboard-Uses-Crew-and-Targets.md: - Fix file references (everything is in R/simulation.R, not separate files) - Remove references to non-existent R/run_async.R, R/run_sync.R - Add correct file list with R/simulation.R as main source Home.md: - Fix dashboard link (.html extension required) All changes verified against actual codebase by parallel agents. Co-Authored-By: Claude Opus 4.5 <noreply@anthropic.com>
fix: Correct algorithm description and deprecation status Key corrections: - Remove "walker steps on existing black pixel" (impossible - walkers terminate when ADJACENT to black, before they can step onto one) - Fix "10-20 messages" claim → actual is O(N) where N = walkers that create black pixels. Broadcasts are sequential, rarely simultaneous. - Fix pseudocode: current position becomes black, NOT next position - Mark Phase 2/3 as DEPRECATED (nanonext sockets fail in crew subprocesses) - Add sync_mode='chunked' as recommended alternative Co-Authored-By: Claude Opus 4.5 <noreply@anthropic.com>
Consolidate wiki: keep only project-specific pages DELETED (generic content moved to llm wiki): - Troubleshooting-FAQ.md (had incorrect cachix commands) - Troubleshooting-Nix-Environment.md - Deploying-Shinylive-Dashboards.md - GitHub-Actions-and-Cachix-Optimization.md - Using-Gemini-CLI-for-Large-Codebases.md - WebR-Performance-and-Single-Thread-Limitation.md - Working-with-Claude-Across-Sessions.md KEPT (randomwalk-specific): - Home.md (updated with links to llm wiki) - Async-Dashboard-Approaches.md - Dynamic-Grid-Broadcasting-Algorithm.md - How-Async-Dashboard-Uses-Crew-and-Targets.md See llm wiki for generic documentation: https://github.com/JohnGavin/llm/wiki
DOCS: Add Dynamic Grid Broadcasting Algorithm wiki page Closes #89 - Comprehensive technical documentation of the minimal broadcasting algorithm New page: Dynamic-Grid-Broadcasting-Algorithm.md - Problem statement (static vs dynamic approaches) - Algorithm overview with pseudocode - Key innovation: minimal broadcasting (~10-20 messages vs millions) - Eventual consistency model with timing diagrams - Implementation details (R/broadcasting.R functions) - Performance characteristics (~12% overhead) - Pub/sub architecture diagrams - Use cases and limitations - Future phases (Phase 1 complete, Phases 2-3 planned) Updated Home.md with new page link
DOCS: Add comprehensive Async Dashboard Approaches comparison Closes #87 - Created new wiki page comparing crew-only vs nanonext broadcasting New page: Async-Dashboard-Approaches.md - Overview of two async implementations - Crew-only (static snapshots) detailed explanation - Nanonext broadcasting (dynamic updates) with architecture diagram - Comprehensive comparison table (12 aspects) - Decision flowchart for choosing approach - Code examples for both approaches - Performance benchmarks Also updated: - Home.md: Added Async & Parallel Processing section - Updated wiki structure diagram
DOCS: Add fractal similarity section (workers=0 vs workers=1) Closes #48 - Explains the architectural differences between sync and async modes, and why workers=1 produces similar results to workers=0 despite using completely different code paths. New section: Understanding Workers: Sync vs Async Architecture - The Three Modes table - Fractal Similarity explanation - Comparison table (7 aspects) - When to use each mode - The 'Fractal' insight
Replace with cross-reference to claude_rix wiki Content migrated to centralized location: https://github.com/JohnGavin/claude_rix/wiki/Nix-Environment-Troubleshooting Addresses issue JohnGavin/statues_named_john#76
Add comprehensive Troubleshooting FAQ - Created Troubleshooting-FAQ.md (comprehensive FAQ page) - 6 main sections: Nix, Cachix, GitHub Actions, R packages, Shinylive, Git - 25+ Q&A pairs covering most common issues - Quick reference section with most-used commands - Cross-references to detailed wiki pages - Updated Home.md - Added Troubleshooting FAQ link (top of Development Environment) - Updated wiki structure - Marked as NEW Covers common issues from: - Nix environment degradation - Cachix authentication and storage - GitHub Actions caching (Magic Nix Cache) - R package development - Shinylive dashboard deployment - Git workflow Related to PR #75, caching optimization work
Add comprehensive caching guide and update related pages - Created GitHub-Actions-and-Cachix-Optimization.md (comprehensive caching guide) - Multi-layer caching architecture (Magic Nix Cache, GH Actions, Cachix) - Performance benchmarks (68% speedup in steady state) - Detailed troubleshooting for all cache layers - Best practices for workflow optimization - Updated Troubleshooting-Nix-Environment.md - Added Cache-Related Issues section - Quick fixes for common cache problems - Cross-reference to caching guide - Updated Home.md - Added link to new caching guide (marked as NEW) - Updated wiki structure - Updated last modified date to December 2024 Related to PR #75 (magic-nix-cache optimization)
Add cross-reference to crew/targets guide
Add comprehensive guide on crew/targets usage in async dashboard Explains with code examples how the async dashboard uses crew and targets: - Crew controller lifecycle and task distribution - Step-by-step code walkthrough (9 steps) - Worker lifecycle visualization - Data flow comparison (sync vs async) - Educational examples (4 basic patterns) - Cross-references to related wiki pages Addresses educational request to explain crew/targets workflow patterns.
Add wiki page explaining WebR single-thread limitation Comprehensive explanation of why WebR uses only one CPU core: - Browser/JavaScript single-thread constraint - Cooperative multitasking vs true parallelism - Performance expectations: ~1.2-1.5x speedup with 2 workers - Comparison with native R (200%+ CPU usage) - Technical deep dive and FAQ Addresses common question about Activity Monitor showing 100% of one core.
Destroyed Welcome (markdown)
Add comprehensive wiki documentation - Home page with navigation - Troubleshooting Nix Environment guide - Working with Claude Across Sessions guide - Using Gemini CLI for Large Codebases guide - Deploying Shinylive Dashboards guide Migrated from repository markdown files for better discoverability.
Initial Home page