Paperhat Limited · Wellington, New Zealand
Rules people can read.
Systems can verify.
Paperhat is developing deterministic semantic specifications for consequential rules and standards. The aim is to preserve governed meaning as it becomes requirements, data, software, tests, guidance, and conformance evidence.
The translation gap
Meaning drifts as rules become systems
Public rules begin as human language. Different teams then translate the same intent into requirements, data models, software, tests, forms, guidance, and assurance material. Each translation can introduce ambiguity, omissions, and divergence.
When an authorised rule changes, teams need to identify what changed, what it affects, and whether the implementation still conforms. Today, much of that knowledge is dispersed across documents, code, tests, and institutional memory.
Deterministic semantic specifications
One governed meaning. Reproducible results.
Explicit meaning
Authoritative language is bound to an explicit semantic representation instead of being repeatedly reinterpreted downstream.
Closed-world validation
Only declared meaning is present. Missing information is exposed rather than silently inferred, and exceptions are represented explicitly.
Source traceability
Interpretations, checks, and derived outputs can be connected to the governed source that authorises them.
Repeatable change assessment
The same governed input is intended to produce the same result, making authorised changes and their effects easier to review.
Current status
An early technical foundation, ready for a bounded test
- Demonstrated
- A scoped bootstrap implementation demonstrates parts of deterministic parsing, validation, lowering, RDF, storage, and querying.
- Under development
- The complete Holotype and Gloss pipeline, generated foundries, and a user-facing product.
- Proposed
- A bounded public-sector conformance trial, 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 is a bounded section of a public technical standard, such as the draft New Zealand API Standard. The trial would test source traceability, ambiguity detection, repeatability, and the effect of authorised changes.
Human decisions stay human
Paperhat does not replace policy authority, legal judgement, standards governance, or professional discretion. Where a source is unclear, the system should expose the ambiguity for an authorised human decision. It should not invent one.
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.