User`s guide

How the Matching Algorithm Works
9-13
• A elements are scored on the ratio of matching B elements
• For both A and B the score is compared with a threshold of 0.5.
The left A and the middle A have two out of three B elements in common, resulting in a
matching score of 2/3=0.66. The XML comparison tool marks the A elements as matched
and the report shows that their contents have been modified.
When a user makes a further change to the middle document (resulting in the right
document), and this new document is compared again to the left document, the matching
score for A drops to 1/3=0.33. The algorithm considers the A elements unmatched this
time. In this case, the difference between the two documents is marked as a deletion of A
from the left document and an insertion of a new A into the right document.
This problem is likely to occur when there is little information available inside a single
element to score a match. A seemingly small change in one of the properties used for
matching could tip the score under the threshold, and therefore result in a large change
in the outcome of the comparison.
Elements Matched Across Different Parts of the Hierarchy
Sometimes unexpected matches of similar items occur across different parts of the
hierarchy. In the following example, C elements are matched on name: