It would be ideal if PerfectMatch could take split or merged segments into consideration. It would especially be useful when you use PerfectMatch to load a new version of a document you are translating. Split/merged segments are not filled because the segmentation re-applies the same patterns as it did by the segmentation of the first document. PerfectMatch could check if the previous version has split/merged segments and make some tests (especially for merged segments) on the new version to see if the content could be re-used at these specific locations