Akoma Ntoso

In short: Akoma Ntoso structures the text of the law; Arxo executes a model of it. There is no execution comparison here by design — only markup and transfer. This page is for publishers and tool builders who move between marked-up acts and pinned sources.

Key differencesChoosing and combiningEvidence & limitations

What Akoma Ntoso is for

Akoma Ntoso (OASIS LegalDocML) is an XML vocabulary for acts and their editions: work, expression, and manifestation levels with mandatory naming, element addressing, lifecycle and temporal groups, and point-in-time URLs for edition slices. Its strengths are maturity as a versioned standard, edition machinery publishers actually serve (structured official text with retrievable slices), and open parsing and publishing tooling.

Where it meets Arxo

The intersection is markup and transfer only: marked document, edition, and fragment against Arxo’s source, edition, publication, and fragment records. The honestly comparable property is round-trip preservability — marked text to pinned bytes to anchors and back — with the publisher’s pinned bytes as arbiter. Amendment semantics and computed output are the canon’s job, explicitly not the markup’s.

Key differences

• Structure is not a rule language. The standard’s own position, as recorded in the study, is that its XML tree is not a canonical rule representation. Asking it to decide cases miscasts the format.
• Pinning granularity differs. Arxo pins section-level editions with validity dates and content hashes; the markup versions whole documents with validity starts and regeneration records. Different granularities, not defects.
• Addressing needs a mapping table. Element identifiers map to article locators only through an explicit table — a naive prefix rule fails by construction, as the protocol’s control case demonstrates. The table is part of the comparison, not an afterthought.
• Exchange gaps are publisher decisions, not standard limits. Nearly every sharp question in the study — vendor extensions, temporal-group counts across slices, elements missing identifiers, regeneration versus retrieval timestamps — turned out to be a property of the publisher’s output. The page attributes them there, never to the standard.

A concrete scenario

The prepared experiment round-trips section pins of a limitation act — share of fragments with found identifiers, exact identifier-to-locator matches, revision preservation across validity dates — plus four edits: a dated slice against the current one including a visible later amendment, whole-act versus section serialization, naive versus table mapping, and a schema-copy control. Converter behavior was never run and is not claimed.

Choosing and combining

Choose Akoma Ntoso when the deliverable is structured official text with addressable editions that other tools can consume. Look to Arxo when the text must additionally execute with dated editions and replayable derivations. Combined, marked-up publications feed pinned sources, and pinned sources anchor executable models — with the mapping table and byte hashes as the joints.

Evidence and open questions

• Sources checked: September 2026 (the vocabulary and naming-convention standards, the frozen schema copy, the publisher’s act bytes read in full with verified hashes, and a European formex reference).
• Studied profile: Akoma Ntoso 1.0 with the XML schema third-edition copy; one publisher’s output as the live exhibit.
• Basis: confirmed by documentation and byte reading plus a prepared protocol with frozen inputs; comparative run not performed — and no executable Arxo query exists for the identifier check, honestly marked as no model rather than stretched.
• Open: publisher-side classes above, one containment question that resolved as a normal outcome, and the run itself.

Sources and reproducible materials

• Vocabulary and naming convention: akn-core-v1.0-os-part1-vocabulary.html
• Publisher exhibit (official act bytes): legislation.gov.uk/ukpga/1980/58/data.akn