Yealink desk phones are everywhere on the second-hand market. Operators retire them in batches, resellers list them by the pallet, and they show up on Marktplaats and eBay for a fraction of the new price. The catch is that a lot of those phones are still administratively “owned” by the previous operator, even after a factory reset. In this post I will explain how I’ve put my own configuration on a set of phones that I bought, without asking Yealink (or the previous owner’s reseller) for permission. This post explains why that is harder than it sounds, and how multicast SIP PnP lets you do it anyway. ...

Generative AI works best when everything is plain text. Code, configuration, documentation, data, the more you can represent as readable text files, the more effectively an AI agent can reason about it and transform it. Database schemas are no exception. You can dump a schema as SQL DDL and hand it to your AI agent, and that works. But SQL DDL is imperative and verbose; it tells you how to create tables, not what the schema looks like. DBML (Database Markup Language) takes a different approach: it is a clean, declarative, database-agnostic language for defining schemas that both humans and AI can read and work with comfortably. ...

I have written a command-line tool in Go called github-export that exports all GitHub issues, pull requests, releases, labels, and milestones from a repository into a local folder as plain markdown files. What makes it different from existing tools is that it syncs incrementally and generates event files that an AI agent (like Claude Code) can pick up and act on. This lets me maintain my open-source projects with AI assistance. ...

For years we have been told that speed is everything. That shipping faster is the only metric that matters. We are being told that the backlog is dead, that human bandwidth is a bottleneck, and that the future belongs to whoever presses the AI’s generate button hardest. We reject this. Not because we fear change, but because we have used these tools, and we have seen what happens when people stop caring about the output. We have seen skills abandoned, standards dropped, and products shipped that no one can explain, maintain, or bring themselves to care about. We have been told this is progress. ...

SQL gives us rows and columns, while APIs usually need nested JSON. That mismatch is small, but it creates a lot of repetitive code. Most backends choose one of two strategies: use a join and transform the resulting rows into nested JSON, or run multiple queries (often one per table) and then match each result set back to its parent. That code is not hard, but it is everywhere. And as Robert C. Martin reminds us in Clean Code, duplication is a design smell, even when each copy is “small”. PathQL is my attempt to remove that duplication while keeping SQL in the center. ...

I wrote phpfilemerger, a small CLI tool that takes a PHP entry point and all of its class dependencies and produces a single self-contained PHP file. It uses an AST parser to resolve dependencies properly, so the output is ordered and ready to run. Use case: PHP-CRUD-API I maintain a PHP project that is distributes as a single file (PHP-CRUD-API) and phpfilemerger automates the process of building. It produces that one-file distribution, keeping class order and namespaces intact so the merged file works just like the original project. It statically analyzes your codebase using PSR-4/PSR-0 mappings and inlines all dependencies in the correct order. ...

I’ve built TQDBProxy to handle heavy read loads by caching SELECT queries with a TTL hint. The proxy worked well for reads. Each application server could keep a local cache and avoid trips to the database. The cache reduced pressure on primary databases and gave predictable latency under heavy read traffic. But what if not the reads, but the writes are the bottleneck? Each small insert or update creates its own transaction and its own fsync on the database. The database CPU and I/O become saturated. Adding more machines does not help as you cannot easily shard hot tables over multiple database servers. ...

I recently evaluated the Cytrence Kiwi Plus, a USB KVM solution that enables remote access to machines via HDMI and USB. The product was provided for free for evaluation. It retails for around 100 dollars. In the past 6 weeks, I’ve tested it daily (every evening) with two real-world scenarios: Linux controlling Mac and Linux controlling Windows. In this blog I will share my extensive experience with the product. It feels just like working on another virtual machine, but this machine is real and on your desk. ...

Recently Patrick Akil (of Beyond Coding) interviewed legendary architect Gregor Hohpe, author of “Enterprise Integration Patterns” and former architect at Google Cloud and AWS. What follows is a masterclass in thoughtful questioning and profound answers. I highly recommend you to watch it. This post lists the top 3 take-aways. 1. The Art of Amplification Patrick’s pointed questions draw out decades of wisdom. When asked what separates great architects from mediocre ones, Gregor’s answer reframes the entire role: ...

In software engineering, thinking is the expensive part. It always has been. Writing code? That’s the cheap bit. Typing fast is great, but don’t let that speed fool you. Here’s something to remember: code is read many times but written only once (ref: Robert C. Martin). The real cost comes later, when you have to fix bad design. That’s when you pay, and you pay a lot. The real work is in the judgment calls, the choices, and the structure you set up at the start. That’s what makes an engineer valuable. ...