Roadmap: make unknown JSON easier to inspect and query

[duriantaco]

Goal

Make jonq the fastest path from unknown JSON to a working extraction command.

The product direction is intentionally narrow: inspect payload shape, find useful paths, write readable jq-powered queries, and export results as terminal-friendly formats.

Evidence so far

• jq is powerful, but many users forget the syntax for common one-off JSON inspection tasks.
• Unknown nested paths are often the real blocker before writing a query.
• Special keys, dotted keys, sparse arrays, and nested arrays create avoidable friction.
• Tools like gron, jello, jq, and DuckDB already cover adjacent needs, so jonq should not become a database or a full jq replacement.

Current bet

Improve the inspect-first workflow:

1. Run jonq data.json.
2. See root shape, discovered paths, types, samples, and copyable query ideas.
3. Use readable select ... from ... if ... syntax for common extraction tasks.
4. Drop to raw jq when the task needs full jq power.

Non-goals for now

• No database engine.
• No joins or window functions.
• No ETL scheduler or connector framework.
• No large flag expansion unless a workflow clearly needs it.
• No TUI until repeated user feedback justifies it.

Feedback wanted

Please open one of the issue templates added in #20 if you have:

• a jq command you wish was easier to remember,
• a JSON file or shape jonq struggles with,
• a real terminal JSON workflow jonq could make smoother.

Small sample inputs and actual commands are more useful than broad feature requests.