• Normalize frequencies in Hz (>10kMhz) to Mhz.
• Remove bogus "zero" fields for A_INDEX, DISTANCE, power, and CQ/ITU zones.
• Improve "source override" selection logic.
• Improve comparing and selecting of text fields like NAME/ADDRESS/STREET/COUNTRY/RIG et al.
• Remove redundant comment if comment is a duplicate of RST fields.
• Select the most "advanced" field response for QSL field processing (e.g. C > Y > N).
• Select longer distance if difference is less that 5 or 15%
• Select frequencies if the difference is less than 10khz.
• Display default values in --help output.
• Remove "empty" items.

"No good deed goes unpunished"

• Encode CSV file as latin-1 as well.
• Include malformed QSOs in the problem-report.json file.

"Be liberal in what you accept, conservative in what you send."

Major changes to ADIF validation.
* Warn and ignore QSO's missing a mandatory field.
* If an Integer or PositiveInteger is a float, convert it to an int.
* Repair common mode/submode mistakes.
* Use locale sane folding instead of lower() when comparing CNTY fields.
* Add --log-level
* Default to latin-1 file encoding per ADIF spec.

• Properly process ADIF Numeric fields as floats, add full ADIF Numeric, Integer, and PositiveInteger sanitization.

Known outstanding issues:
UTF-8 vs cp1232 vs ISO-8859-1 vs ISO-8859-15 encoding differences on input .ADI file formats.
The specification demands that they are in plain ASCII which nobody follows.