The structure of the xml file to be processed in ReaderBench is as follows (UTF-8 encoding):
<Dialog team="1" file="file-name.xml">
It is easy to use a spreadsheet file to semi-automatically populate a file (see the attached file as an example).discussion-example-RB.csv.
The "nicknames" column shows the participants of the discussion, the "utterance" column show their utterances.
Then, use a word processor to search/replace every $ sign by a return carriage, and every € sign by a tabulation.
Eventually, add the first and latest codes (Dialog, participants, etc.).