Hi Andrew, Thanks for stumbling into this problem. We can confirm that the UniHan readings are incorrect, and will generate a correct mapping between the 164 CJK characters in question and their kVietnamese values for submission to the UTC.
Best, James
