Hi Clara,

The PEM percentage can be thought of as the percentage of similarity between the previous (original) and updated target; calculated after a PE operation has taken place; used to understand the effort during PE phases. It is simply: 100 – (( Edit distance / Relative distance)*100)

The edit distance is the minimum number of edit operations required to transform the one string into another; the comparison in this case is between the previous (original) and updated target.

To understand the how to calculate the percentage of changes, we would need to know a relative distance; in this case, that is the number of characters in the Target. It takes the value from the original or updated target, depending on which is greater.

A record is created each time a segment is updated, which includes a lot of useful data, e.g. time the user entered and exited the segment, the original and updated states of various properties, edit distance, etc…

As you pointed out, it is possible that the same segment is visited and updated multiple times. To understand the edit distance of the overall changes that occurred, you would need to recalculate the edit distance taking the first state of the original target and comparing it against the last state of the updated target.  This recalculated edit distance is included with the reports that are generated by Qualitivity but not maintained physically with the records in the database and/or currently exported with the raw data; simply because it is recalculated given the input used in the report.  

However, it should be possible to include this data in the export formats (excel, xml); would require an enhancement.

Patrick