Questions like "遇到农历闰月生日，怎么提问才不把历法输错" look like minor technique questions, but they decide whether the whole later reading will drift. If 历法对齐 and 原始出生记录 are not set correctly first, the answer often follows mood and approximation instead of structure.
Birth-time boundaries, calendar conversion, and asking order are not small details. Once the entry point is corrected, the same topic usually becomes far more specific and testable.
Set the input and the asking order correctly first
The key move is not asking one extra question. It is setting 历法对齐 and 原始出生记录 correctly before the reading opens up.
Once the entry is stable, you can tell whether later vagueness comes from missing data or from a question that is still too broad.
Where the question most easily drifts
The steadier move is to preserve the original family record first: which lunar year, which leap month, which day, and whether it was daytime or night, then let the system convert.
Many leap-month errors do not come from the reading itself. They begin with the wrong calendar assumption at input.
Do not rush to the conclusion first
Do not treat calibration as wasted turns. Many empty follow-ups come from a weak first setup.
And do not rush toward the result first. If the entry is unstable, more result-chasing usually creates more error.
A steadier asking order
- Shrink the input range until it becomes comparable.
- Use one or two easy verification points to test the difference.
- Only then move into the full follow-up.