Older articles.

How to pin down the behaviour of a legacy system before you change it: characterization tests, golden master, seams, and when to stop.

Read the article

How I assess a legacy system in five days before somebody buys or inherits it: the order of work, the commands, and what ends up in the report.

Read the article

Modernization is not always the right answer to a legacy system. When freezing, standard software or retirement is the better call, and how to tell.

Read the article

When the developer of a running system is gone, the first week decides the cost. What to secure before touching the code, in the right order.

Read the article

Whether you must issue e-invoices from January 2027 depends on your 2026 revenue. What counts, what is exempt, and why receiving them is not enough.

Read the article

How I moved my knowledge into a Git repository so agents can work with it. What held up after nine days, what did not, and what a team can take from it.

Read the article

Five major versions in one jump. Loud errors are cheap. The dangerous break is silent, and your customer finds it first. What breaks, how I run it, what it costs.

Read the article

No security fix for almost ten years, and it still carries your revenue. Why zf1-future buys you time but not support, and the three honest ways out.

Read the article

Twelve months of FinOps across two AWS accounts: where the large amounts were and why commitments come last.

Read the article