Blog

This website ran for years on a very old JBake fork and has now moved to Hugo: 73 articles in two languages, old URLs, audio, admonitions. Implemented with an AI agent, steered through a Werkbaum. This article describes the process and why Werkbaum proved itself as a planning tool for agentic AI in particular.

SQL and CSS are both declarative languages. Why do many backend developers write complex SQL queries with ease, yet despair over CSS? A comparison of locality, hidden dependencies, and the question of how declarative both languages really are.

Seitdem ich, vor langer Zeit, mit dem Konzept Secrets-Manager in Kontakt gekommen bin, hatte ich ehrlich gesagt immer nur eine vage Vorstellung davon, was für einen Vorteil diese eigentlich bieten, wo doch das System, welches ein Secret abfragen darf, sich dafür bereits authentifizieren muss. Wie genau autorisiert sich also z.B. ein Webserver gegenüber einem Secrets-Manager, um dort Datenbank-Zugangsdaten abzurufen, und …

This article examines Lombok’s val — how it uses annotation processing to infer local variable types and enforce finality—highlighting benefits like cleaner, more readable code, compile-time safety, immutability, pre–Java 10 support, and seamless Lombok integration, while weighing drawbacks such as redundancy with Java 10+ var, reduced flexibility due to final variables, reliance on compiler internals with potential compatibility risks, and possible lag in new JDK support; it recommends using val chiefly in Lombok-heavy, pre–Java 10, immutable contexts, but preferring native var in modern Java and planning a long-term shift toward Java-native features.

While most AI code generation examples show how to quickly bootstrap a new web application, I was curious how far an AI coding assistant can take me in an existing, non-trivial codebase – one with a Java backend, automated tests, APIs (OpenAPI), PostgreSQL database migrations (Liquibase). So, I decided to explore the open-source AI tool aider, applying it to one of my Open Source projects.