Chapter 2 of 11 · 2 min read
Five Architectural Invariants
The short rules that keep a shared visual language from turning into shared authority or shared disclosure.
The ~WellbeingIdentity Rendering Architecture · v0.2 · edition history · Moses Sam Paul
Quick Answer
The rendering architecture is held together by five rules: one canonical identity anchor; projection is not record; rendering is not authority; Pilot W is not canonical WellbeingScore; and Performance is not wellbeing state. They are design constraints, not marketing claims.
The release began with a simple observation: a shared portrait could make several products look coherent while quietly blurring the lines that make the system safe. The answer was not more UI. It was a short set of invariants that every interface, package, and public explanation must respect.
- 01One canonical identity anchor
A stable identifier and shared node semantics prevent each surface from inventing a different person.
- 02Projection is not record
A purpose-limited view is assembled for a job; it is not a private record with fields merely hidden from sight.
- 03Rendering is not authority
The renderer presents supplied data and emits interaction events. It does not decide identity, consent, disclosure, or calculation.
- 04Pilot W is not canonical WellbeingScore
A live pilot signal remains experimental; clearer presentation cannot turn it into a validated composite measure.
- 05Performance is not wellbeing state
Outward contribution or performance routes remain distinct from the identity-state reading.
Why invariants matter
The first two rules protect meaning and minimisation together. A stable anchor prevents accidental identity duplication; a projection boundary prevents that anchor becoming an excuse for indiscriminate access. The third protects governance: a component can be reused only because it has no discretion to reach behind the supplied projection.
The last two protect interpretation. A score-looking number can acquire authority through repeated display long before its methodological status has changed. Likewise, an outward route can look like a personal state when colour, position, or motion collapses the distinction. The design explicitly resists both shortcuts.
A test for every new surface
Before a surface is added, its owners should be able to answer five questions in plain language: What identity anchor does it use? What projection purpose is it receiving? Which authority created that projection? Which values are experimental or separate? And which visual choices might falsely imply more authority than the system has?
If those answers are vague, a new component is premature. If they are precise, a shared visual grammar can improve recognition without expanding access.