From a Reliable Source
Uit Betrouwbare Bron (“From a Reliable Source”) is a research and design project by the VNG (Association of Dutch Municipalities), funded by the Ministry of the Interior and Kingdom Relations (MinBZK), that explores a fundamentally different approach to government data registries. The central question: how do you design registries so that the provenance, meaning and consequences of data remain transparent and — where needed — correctable?
Core problem
Data travels fast through administrative chains. A correction made at the source rarely finds its way back through the entire chain. The result: inconsistent records, citizens stuck on errors they cannot fix themselves, and government organisations unaware of how reliable the data is on which they base their decisions.
The project’s starting point is epistemic humility: modesty about what the government thinks it knows. Registries do not contain facts — they contain registrations of observations, decisions, at a specific moment, by a specific party, with a specific degree of certainty.
Two workstreams
The project works along two parallel lines:
- A practical guide for reliable registries — a usable reference for designers and architects of government registries, with concrete principles and patterns.
- Research into Atomic Claims & Annotations (ACA) — an exploration of a technical model that allows individual data elements to carry provenance, validity and certainty, independently of the registry in which they reside.
Five architecture-defining aspects
The guide is structured around five architecture-defining aspects:
- Provenance — where does a data item come from, who registered it and on what basis?
- Registration — how and when was a data item recorded in the registry?
- Validity — for which period or context is a data item valid?
- Certainty — to what degree is the data item reliable, and what is the basis for that certainty?
- Deregistration — how and when does a data item leave the registry, and what are the consequences?
Background
The project builds on years of practical experience with major Dutch base registries: the BRP (population register), BAG (addresses and buildings), BRK (land registry), WOZ (property valuation) and the UWV Polisadministratie (employment and benefits administration). That experience makes clear where things structurally go wrong — and points towards a better approach.
My role
I participate as solution architect. My specific contribution is adding and elaborating Event Sourcing as an architectural pattern in the guide. Event Sourcing makes it possible to capture the full history of a registry as a sequence of immutable events — so that provenance and correctability are not exceptions, but inherent properties of the system.
I also act as a hands-on contributor: not just thinking about direction, but helping produce a usable, concrete end result.