Problem Paper Contributions#

design-research-problems can translate a packaged problem into deterministic Background and Methods support for a later paper draft. The export preserves problem, prompt, variant, family, study, and citation provenance without claiming that the configured task was executed.

Collect a Component Packet#

Use the stable catalog identifier:

import design_research_problems as drp

packet = drp.collect_problem_paper_contributions(
    "ideation_peanut_shelling"
)

The returned JSON-compatible object follows paper-contribution contract version 0.1.0, accepted by design-research-experiments. It contains:

  • configured Background and Methods contributions;

  • curated references deduplicated by the existing Citation.key values;

  • prompt, variant, family-lineage, and linked-study identifiers when available;

  • exact ideation prompt wording and its evidence tier; and

  • reporting gaps for provisional citations or incomplete linked records.

Evidence Boundary#

Every Problems contribution has evidence_basis="configured" and an empty evidence_refs list. A catalog label such as human-subjects-ready is a task-selection property, not proof of participant recruitment, execution, or analysis. The Methods metadata therefore reminds the downstream renderer to obtain actual participant, timing, assignment, evaluator, and modification details from study evidence.

Ideation Provenance#

For a problem linked to the ideation catalog, the helper follows the existing source_citation_keys and prompt relationships. It exports the selected prompt, its source-specific variants, its family and ancestor lineage, and any study records that directly reference those variants. It does not broaden the selection to sibling prompts merely because they share a family.

The prompt contribution includes the complete packaged Markdown statement, including its constraints and instructions. Its metadata retains that exact text as statement_markdown and labels the shorter ideation-catalog brief as catalog_summary; the summary is not a substitute for the selected prompt.

Non-Ideation Problems#

Decision, optimization, grammar, MCP, and ordinary text problems use the same packet shape. They receive general Background and Methods contributions from ProblemMetadata plus their curated problem citations. Decision and optimization metadata also carries a reminder to report the variables, objective, constraints, and evaluator settings actually used.

Downstream Aggregation#

Pass the packet to the Experiments collector as a component packet. Experiments validates citation keys, deduplicates compatible references, and keeps the Problems package and version as provenance. The helper generates neither a manuscript nor observed Results claims.

See examples/catalog/paper_contributions.py for a runnable example.