Feature/schema decision troubleshooting

What is the learning goal?

Use HVE to decide what features are needed and what schema supports them. Technology capabilities matter, but coding, formula/debugging and designer configuration are not the core task.

How is it technology-specific without coding?

Canvas offline/media/checklist choices, model-driven relational casework, flow events/follow-up, agent knowledge/actions, Pages self-service/visibility and BI grain/lineage have different feature/data implications. The learner must explain those implications, not just recite product names.

HVE starts by giving a schema

Ask which client need and selected feature justifies every entity/field/relationship. Request a smaller alternative. Do not treat the supplied CSV or earlier maker material as a database blueprint.

Does copying the prompt also copy its evidence?

Yes. Copy standalone prompt includes the task, decision-step definitions, complete interview, applicable RULE text and raw examples. Expand the complete-prompt section to inspect it. The Markdown download contains the same payload; no course links or attachments must be read to obtain this static context.

Later steps still require real learner choices or verified earlier outputs. The prompt explicitly asks for those if missing and pauses dependent work. Static evidence cannot supply an actual approval, preferred model, run ID or runtime observation.

HVE lists every possible feature

Require first release, later, rejected or needs validation choices, with value/evidence/dependencies. A product capability is not automatically a business requirement.

HVE jumps into code or implementation repair

Ask which unresolved feature/data decision the implementation would answer, and return to that choice. Implementation can follow an accepted design; it is not proof the right design was chosen.

How do I judge the proposed schema?

Trace need -> feature -> data. Walk realistic records. Check zero/multiple children, current state versus history, required/optional values, ownership/access and changed requirements. Challenge both missing data and unjustified complexity.

What if two models both work?

Record the simpler model, richer model, tradeoff and assumptions that would change the choice. HVE should help you make an informed decision, not pretend there is one universal table count.

Where is the correct output?

Use the selected project's START-HERE.md available-deliverable link or library/deliverables.md. Match actual run/artifact/version and substantive content. Follow the output guide for missing links, stale mirrors, older labels and read-back-only work.

Does selecting a model approve the server gate?

No. Workshop planning disposition, server review/acceptance, mirror state and operator approval are separate. Editing a worksheet row does not release a gate.

What do I do after reading the output?

Perform the guide's Find/Do/Next task: correct a need, choose features, compare a model, walk sample records or decide a scope change. Record the actual artifact and reason, then carry that decision to the next request.

What if sources or the plugin are unavailable?

Report the actual blocker. Do not invent file contents, run IDs or original artifacts. A desk walkthrough is useful but is not live HVE practice. Use normal same-run questions/status and authorized operator handoffs; do not bypass through specialist tools.

Are the review findings actual app tests?

No. They are fictional design walkthroughs. Actual HVE outputs and any actual application observations are separate evidence. The current interview pack is HVE-DECISIONS-1.0.

Does HVE lack implementation capability?

No such general claim is made. Capabilities depend on the host/tools and configuration. These exercises deliberately focus on design decisions. A decision artifact by itself does not prove implementation happened.

What happens after the decision pack?

Use it for client design review and prototype/build planning. Earlier maker references remain optional downstream material, not prescribed schemas. Actual implementation needs its own authorization, inputs and evidence.

Is local validation workshop certification?

No. It checks the authored pages and links. Live HVE rehearsal, representative learner validation and timing remain separate. Public release requires explicit approval.