A Business Backend for My IDE
summary
An LSP-based backend built in Pharo exposes model-driven analyses to IDEs by combining Moose Technology (Famix/Fast metamodels), the Motion pattern language, and GitHub’s incremental parser (Tree‑sitter/Triciter) to generate models and run rules. It supports very large codebases (reported 100M LOC imports in ~1–1.5 hours) with an incremental model updater that produces live checks in ~400–700 ms per file change, and demos show VS Code/IntelliJ plugins enforcing comments, method/line limits and architectural layer rules; the system runs locally, is open source, and is used for microservice splitting, architecture verification and modernization.
AI-written summary · about
chapters (11)
- 0:00Introduction
- 1:33Code editor limits
- 5:44Model-driven engineering
- 9:29IDM use cases
- 12:37IDM in IDEs
- 16:07LSP protocol
- 20:23Real-time analysis
- 26:43Architecture violation demo
- 31:14Error handling trust
- 34:01Academic open source
- 36:27Conclusion
related talks
40mIntegrating IDEs with Dotty the experimental Scala compiler
46mSix steps from zero to IDE
37mTextual and Graphical Languages for the Cloud Era
39mIDE development with Ruby
36mCode as Data: How data insights on legacy codebases fill the knowledge gap in modernization projects
43mCode quality as an engine of your IT business
29mExploring Scala Tooling ecosystem
23mCode smarter, not harder: How AI Tools Boost Your Productivity