Heise is reporting on RUML, a file format designed to translate laws from prose into machine-readable logic.

The pitch is straightforward: legislation is currently written for people, which means every piece of software that has to obey it, tax calculators, licensing systems, benefit eligibility tools, and increasingly AI agents, must reinvent an interpretation of the same text, and do so differently each time. A format that encodes provisions as structured logic would let a rule be evaluated rather than paraphrased.

That puts RUML in a long line of attempts to make law computable. Two decades of work on standards such as LegalXML and the LKIF ontology have shown the hard part is not the schema. It is capturing qualifiers such as reasonable, without undue delay, and in the public interest, which were deliberately left open for humans to argue about. A logic file forces that argument into the open, which is either the point or the problem, depending on who is reading.

The potential payoff is real. Machine-checkable rules would let regulators publish rules that can be tested against themselves, let companies run compliance checks before shipping rather than after, and give AI systems a canonical source instead of a PDF they summarise.

What remains open is who stands behind the format, how far adoption has progressed, and whether it is being positioned as a standard or a proposal. Those questions decide whether this is a specification or a footnote.