Comment by ChrisSD

In summary the two useful measures of unicode length are:

* Number of native (preferably UTF-8) code units

* Number of extended grapheme clusters

The first is useful for basic string operations. The second is good for telling you what a user would consider a "character".