In order to improve translation memory, it should be easy to get real test data from different production environments.
We need a new scripts to export and import this data. The format of the data should be different, more generic than what is currently stored in the schema. It should be possible to do a schema change in TTMServer index and just use the same dump to re-import the data.