FAQScope, sources and ownership
Is the schema fixed, a reference, or tailored?
The public Blueprint schema is the reference starting point. InPerceptic offers schema tailoring today as a scoped service when your questions reveal a gap. Current tooling reads YAML. New concepts or notation can need validation, extraction, graph, query or rendering adaptations; adapters are priced separately.
Do we have to model the whole system?
Start with one change initiative and one bounded context. Select the knowledge needed for the agreed questions, and record what is outside the boundary. Expand when another decision justifies it.
Who authors the model, and who reads it?
Named authors and reviewers maintain the YAML and its source links. Other team members can use the agreed rendered views without authoring YAML. AI agents can query a model where the selected tooling exposes that access; access and review responsibilities are part of scoping. The quality of people's and agents' answers still depends on the model's coverage and review.
Can several models refer to the same code?
Yes, models can record different questions about shared code. Give each model an explicit scope and map its references to the relevant repository, path and revision. Shared code or a shared namespace does not establish automatic semantic federation between schemas.
Can one model span repositories?
It can record a system that spans repositories when the source mappings are explicit. Agree which repositories, revisions and owners each reference means, and review those links as code changes. Colocation and identifiers alone do not reconcile the sources.
Does a checked model prove the implementation is correct?
No. Record whether a statement is intended design or observed implementation. Link authoritative code, automated tests and runtime evidence, retain their review dates, and reconcile discrepancies. Model checks cover specified constraints. A missing test link is a traceability gap, not a test result or runtime coverage measure. Examine modeled change impact, then verify the implementation separately. Check each rendered artifact against the accepted model revision. A consistent view does not establish runtime correctness.
What does this add to diagrams and a wiki?
Typed relationships can connect a rule to its operation, rationale and evidence, then be reused in different views and queries. Keep explanatory prose and workshop diagrams where they help. Review the model and rendered views alongside the sources; no documentation format maintains itself.