A Java Geek weekly 144

The need for shorter and shorter Time-To-Market requires to integrate more and more third-party libraries. There’s no time for the NIH syndrom anymore if it ever was. While most of the time, the library’s API is ready to use, it happens that one needs to "adapt" it to the codebase sometimes. How easy the adaptation is depends a lot on the language.

Nothing really, to be honest.

Why AT URIs are not valid under IETF RFC-3986, what the impact is, and some possible ways forward.

Tailscale Funnel lets you route traffic from the broader internet to a local service running on a device in your Tailscale network (known as a tailnet). You can use it to share a local service, like a web app, for anyone to access—even if they don’t use Tailscale.

Today I’m kicking off my most ambitious Maker’s Pet project yet: oomwoo, an open-source home robot vacuum that you can build yourself. Open hardware, open firmware, open software — and built in public, from the first commit.

No cloud required. No vendor lock-in. It maps your home with an affordable 2D LiDAR and navigates on its own, runs locally, and integrates natively with Home Assistant. If you’re into Raspberry Pi, ROS 2, 3D printing, or just the idea of owning a vacuum you fully understand and control — this one’s for you.

The fact that non-reviewed code can be tested or deployed in production is surprisingly the most significant benefit. We do not ever block the flow of work through the value stream. We can already obtain valuable feedback from testing, and production. We do not have to wait for a reviewer, or worse the ping-pong between reviewer and reviewee, for the feature to be available for testing or deploying into production.

We want to be explicit about what this study does not establish. This is an observational study, not a randomized experiment, and we cannot make definitive causal claims. Teams that choose to migrate to Kotlin may differ from those that stay on Java in ways we cannot fully observe − though our validity checks (JDK-version segmentation, multiple control groups, stepwise confounder elimination) suggest these differences alone do not explain the gap. We encourage readers to examine the limitations in Section 4.4 and the open questions in Section 4.5 when forming their own assessment of the evidence.

• Safety: Kani automatically checks for many kinds of undefined behavior. This makes it particularly useful for verifying unsafe code blocks in Rust, where the “unsafe superpowers” are unchecked by the compiler.
• Correctness: Kani automatically checks panics (e.g. unwrap() on None), arithmetic overflows, and custom correctness properties, either in the form of assertions (assert!(…​)) or function contracts.

The significant factor to observe here is the competitive advantage, not technical specifics. For two years, the belief was that owning the best model meant controlling the developer.

This perspective is now shifting. Replacing Claude with GLM or Codex, gstack continues to operate because the core intelligence evolved, but the documentation did not.