I’ve restructured a previous pre-print into two different papers. The first focuses on cataloguing calibration in popular semantic parsing systems, and the second looks at what we can do with a well-calibrated model.