Paperhat Limited · Wellington, New Zealand

Meaning specified once.
Preserved everywhere.

Paperhat is developing a deterministic, closed-world semantic system in which meaning is specified once in a governed source. Requirements, data, software, tests, guidance, and conformance evidence are derived from that source rather than interpreted again.

The translation gap

Every interpretation can change the meaning

Consequential public rules and standards begin as human language. Different teams then interpret the same source as requirements, data models, software, tests, forms, guidance, and assurance material. Each independent interpretation can introduce ambiguity, omissions, and divergence.

When an authorised source changes, teams need to identify what changed, what it affects, and whether every downstream form still preserves the specified meaning. Today, much of that knowledge is dispersed across documents, code, tests, and institutional memory.

Deterministic semantic specifications

One source governs every projection

One governed source

Meaning is specified explicitly in a canonical Holotype source that remains authoritative across every downstream form.

Closed-world validation

Only declared meaning is present. Missing information is exposed rather than silently inferred, and exceptions are represented explicitly.

Faithful projections

Requirements, graphs, software, tests, guidance, and evidence are derived from the governed source rather than independently reinterpreted.

Authorised change

An accepted decision changes the canonical source; affected projections can then be regenerated and assessed against it.

Current status

An early technical foundation, ready for a bounded test

Demonstrated
Parts of deterministic parsing, validation, lowering, RDF, storage, and querying have been exercised in a scoped bootstrap implementation.
Under development
The complete Holotype and Gloss pipeline, generated foundries, and a user-facing product.
Proposed
A bounded public-sector trial of declared-design evidence, defined with a suitable agency or delivery partner and evaluated against agreed success criteria.

First external validation

Start with one small, consequential rule set

The proposed first validation would examine declared-design evidence against a bounded section of a public technical standard, such as the draft New Zealand API Standard. It would test source traceability, ambiguity detection, repeatability, and the effect of authorised changes—not the behaviour of a running service.

Read the proposed trial →

Human decisions stay human

Paperhat does not replace policy authority, legal judgement, standards governance, or professional discretion. Where a source is unclear, the system exposes the ambiguity for an authorised human decision rather than inventing one. An accepted decision enters the governed source; affected projections are then regenerated from it.

Help define a credible first test

Paperhat is seeking public agencies, standards owners, delivery teams, and mission-aligned partners willing to explore a safe, useful validation.

Contact Paperhat →