Unit 7: Startup Recovery and the Java-to-ROS Boundary

By now you can create a workspace, build a bringup package, install a launch file, and verify the launched graph. Week 16 asks whether you can recover when one layer is wrong without destroying evidence or pretending two separate runtimes are integrated.

Chapter question

How can you isolate startup failures by source, build/install, shell, launch, and graph layer, then document a repeatable ROS bringup without overstating the Java/WPILib integration state?

What you will understand

You will learn clean-shell diagnosis, package-prefix verification, stale installed-resource failures, missing runtime dependencies, post-launch graph mismatches, startup runbook design, and the architecture contract required before a Java/WPILib application can be claimed to exchange data with ROS 2.

What you will do

You will diagnose wrong overlays, missing launch resources, stale installed launch files, missing dependencies, one-sided remaps, and unsupported integration claims. Then you will write a clean-shell runbook and a separate Java-to-ROS Integration Contract.

What you will produce

Your final artifacts are a Repeatable Startup Runbook and a Java-to-ROS Integration Contract showing what the ROS bringup owns, what the Java/WPILib runtime owns, what a future bridge must own, and what evidence would be required to prove data crossed the boundary.

How the week connects

Weeks 14–15 built the workspace and launch system. Week 16 turns that infrastructure into an operational practice another student can reproduce. Unit 8 will use the complete Java/WPILib and ROS evidence stack to perform a final system-readiness review and capstone defense.

Key vocabulary

clean shell, package prefix, stale install, runtime dependency, fault isolation, runbook, integration contract, bridge, adapter, source of truth, freshness, startup dependency, verification chain

Approximate time

About 210 minutes for the five lessons, lab, and wrap-up. The failure cases can be completed with supplied evidence; students with a local Jazzy workspace may reproduce the safe package/launch failures locally.

Final evidence required

A tested startup runbook that begins from a clean shell, verifies the intended workspace and graph, contains bounded fault-isolation steps, and clearly states that ROS bringup success does not prove Java/WPILib integration without an explicit bridge, data contract, and cross-runtime evidence chain.