Long-lived reference material that supports Tapestry architecture, design, implementation, and operations, including conceptual comparisons, deployment notes, and usage-oriented docs.
| Document | Description |
|---|---|
glossary.md | Definitions of Tapestry-specific terms (consortium training, Shared-Base Loop, Sovereign Build, etc.); links to the canonical ADR for each |
ARCHITECTURE.md | Single-page synthesis of Tapestry's architecture (core-plus-sovereign, consortium training, the training loop, the sovereign pipeline, data sovereignty, and the design invariants), with links to the source-of-truth ADRs |
training-approaches.md | Centralized vs. federated vs. consortium training — comparison table, shared consortium-training-loop.svg, links to ADRs |
training-pipeline-data.md | LLM training pipeline stages, data type taxonomy, quality framework, and implications for Tapestry's consortium model |
This material summarizes the design chain under ../architecture/ and the decision records under ../architecture/decisions/. Where this reference and an ADR disagree, the ADR is authoritative. (Please file an issue for any discrepancies!)