The following behavior was discovered:

• Editing a segment with a written HTML Tag ("<FilesMatch>" in our default import example) is removed by server side.
  → This is wrong and should be fixed, so that the content remains with html entities
• A newly entered Tag <FilesMatch> is saved as HTML entities to the server, but is then removed also.
• If we enter a <b> bold tag, on Marcs chrome it is escaped, in Thomas firefox instead it is not removed by the server but returned as plain HTML <b> tag.
  → This is also wrong, it should be also returned as HTML entities.

This leads to the next Problem:
Since in our Segments Table our content and mqm tags, and terminology resides are stored as plain HTML, all user data must be stored as html entities.
This must be ensured in our application for all steps: import, editing and export! For import and export data storage depends again on file type. While in XML formats we assume HTML entities already as entities, in CSV they might be plain text!

For TermTagging this implies:
Since we send also HTML Tags (our content tags, and so on) to the TermTagger, the termtagger must recognize HTML Entities in Terms and convert them back to the used chars.