Three Cloud Look-Alike Pairs, Six Reliable Tells

Goal Create a compact field dataset that helps a careful reader distinguish three pairs of commonly confused clouds using stable, authoritative references. Deliverable Submit exactly one UTF-8 CSV file named cloud_lookalikes.csv. It must contain the exact header below plus exactly 3 data rows and be no larger than 12 KB. cloud_a,cloud_b,shared_visual_trait,decisive_difference,typical_height_a,typical_height_b,weather_clue,source_a_url,source_b_url Observable requirements 1. Cover exactly 3 distinct look-alike pairs and exactly 6 different WMO-recognized cloud genera, species, or varieties; do not pair a cloud with itself or reuse a cloud. 2. Each pair must be visually plausible. shared_visual_trait must contain 8-20 original words. 3. decisive_difference must contain 18-40 original words and describe an observable, non-circular distinction. 4. Each height field must use a qualitative or approximate band. Include units when using numbers and acknowledge regional variation rather than giving false precision. 5. weather_clue must contain 12-30 words and describe a cautious association, not a prediction or safety instruction. 6. Provide one working, directly relevant source URL per cloud from WMO, NOAA, the Met Office, another national meteorological service, or a university atmospheric-science source. 7. Use accepted cloud names and stable facts. All CSV fields containing commas must be correctly quoted. Negative scope No current forecasts, storm-chasing or safety advice, copied source prose, Wikipedia or listicle sources, personal data, accounts, purchases, outreach, fabricated citations, or additional files. Pass/fail acceptance gates - Exact filename, exact schema, exactly 3 data rows, and no extra files. - Exactly 6 unique qualifying cloud names across 3 plausible pairs. - Every source URL works, is authoritative, and supports the named cloud or claimed distinction. - Every bounded text field meets its word limit and the CSV parses cleanly. - No prohibited content from the negative scope. Scoring rubric (100 points, in order) 1. Factual accuracy and source relevance: 45 points. 2. Quality of pair selection and decisive visual tells: 25 points. 3. Format, count, and instruction compliance: 20 points. 4. Clarity and field usefulness: 10 points. Tie-breaker Prefer the submission whose names, distinctions, height bands, and sources can be verified most directly; then prefer clearer phrasing. Creative pair selection is welcome outside the mandatory core.