.png&w=256&q=75)
1
1
Looking for experience!

Ripple is a cross-repository code impact analyzer built on IBM Bob Shell. It solves one of the most expensive problems in software development: understanding the downstream consequences of a code change before it breaks production. Tools like GitHub Copilot and Cursor answer "what does this file do?" Ripple answers something fundamentally different — "if I change this, what breaks everywhere else?" This is only possible because IBM Bob holds full repository context, not just a single file. How it works: A developer describes an intended change — for example, "rename field lastName to surname in Owner.java." Ripple's Spring Boot backend calls IBM Bob Shell as a subprocess, passing the full repository context. Bob reads the actual codebase, traces every reference, dependency, and downstream usage, then returns a structured impact map listing every affected file, the reason it is affected, and a risk level: HIGH, MEDIUM, or LOW. We demonstrated this on Spring PetClinic, a canonical Java Spring Boot codebase. A single field rename in Owner.java caused Bob to identify 22 affected files across 6 modules — Java service classes, repository interfaces, HTML templates, SQL schema files, i18n properties in 8 languages, and JMeter test plans. Ripple exposes two APIs: POST /api/analyze triggers a live Bob analysis and returns results with a unique ID. GET /api/analysis/{id} lets the frontend fetch results asynchronously. A GitHub Action runs Bob automatically on every pull request and posts an impact summary as a PR comment — giving teams instant visibility before code is merged. The frontend renders an interactive ripple map — a visual graph where nodes are files and edges show dependencies, with HIGH risk files highlighted in red. Ripple was built in 48 hours by a team of 6. It demonstrates Bob's unique repo-context superpower in a practical, shippable developer tool with real business value.
17 May 2026