Text edits example
Text edits are based on the Levenshtein algorithm. It “observes” how many keystrokes are required to transform an initial text into the changed text. If a new text has 10 more characters it will assume an edit distance of 10. For a more precise and complete description see the Wikipedia page: Levenshtein algoritm.
The following is an example to demonstrate how the edit distance properties are calculated in the Text edits report .
Segment | Initial text | Updated text | Edit distance (ED) | Max length (L) | Normalized ED |
---|---|---|---|---|---|
1 | abcd | 4 chars | 4 chars (maximum of 0 and 4) | 1.00 (4 divided by 4) | |
2 | abcd | abcdef | 2 chars | 6 chars (maximum of 4 and 6) | 0.29 (2 divided by 6) |
3 | abcd | abcd | 0 chars | 4 chars | 0.00 (0 divided by 4) |
Given these figures, we now have the total for all the changes:
editDistanceSumLength= 16 : It is the sum of all L values 4 + 6 + 4
editDistanceSum = 6 : Sum of all ED values 4 + 2 + 0
editDistance = 0.38 : Is 6 divided by 16 rounded up to the 2nd decimal.
editDistanceSumNorm = 1.34. It is the sum of individual ED / L fractions. Calculation is (4 / 4) + (2 / 6) + (0 / 4)
Note that normalized edit distances are always rounded up to the 2nd decimal.