User`s guide

How the Matching Algorithm Works
9-15
Two Sequences of Elements Are Cross-Matched
It is difficult to distinguish many similar potential matches and this could produce
unexpected results. In the following example, B elements are scored on name, p1, and p2,
and the score is compared to a threshold of 0.5.
The right document contains one B element more than the left document, and therefore
one of the B elements on the right must remain unmatched and the tool will mark one as
inserted. However, since most B elements on the left potentially match most B elements
on the right, it is impossible to predict exactly how the sequences will be matched. For
instance, the comparison could generate the following result:
“B name=”1” …” > “B name=”2” …”
“B name=”2” …” > “B name=”new” …”
“B name=”3” …” > “B name=”3” …”
“B name=”4” …” > “B name=”4” …”
In this case, “B name= “1” on the right remains unmatched. As in the previous example,
this depends on how all of the other B elements in the two documents are matched. This
situation is likely to occur when elements have several potential matching candidates.