Source: Mark Davis
Date: March 19, 2012
Subject: Process for kMandarin and kTotalStrokes
We originally built the kMandarin and kTotalStrokes by leveraging the existing mappings. At this point, they are not generated anymore. Fixes or additions can certainly be made, however:
  1. They must follow the documented requirements, also listed below, and for any change we need evidence that the new form is the most commonly used.
  2. They must be more carefully reviewed than changes in, say, kDefinition, because they will change sort order.  In particular, any changes must be surfaced for broader review.
We meed to discuss what kind of process is as light-weight as possible for the maintainers of Unihan, but meets the requirements above. One possibility is that (a) we have a new forum for changes to these fields, where proposals can be discussed, and made if it looks like there is rough consensus, and (b) the UTC be notified of all the changes before release.

kMandarin: The most customary pinyin reading for this character; that is, the reading most commonly used in modern text, with some preference given to readings most likely to be in sorted lists. 

kTotalStrokes: The total number of strokes in the character (including the radical), that is, the stroke count most commonly associated with the character in modern text using customary fonts.
Both: Multiple Value Order: When there are two values, then the first is preferred for zh-Hans (CN) and the second is preferred for zh-Hant (TW). When there is only one value, it is appropriate for both.