When loading data containing characters from another alphabeth (for example cyrrilic alphapeth), the data is not imported correctly. In such a case you have to change your Windows language setting to the correspdonding Region/Country.
Therefore, I would be useful to have Unicode support in Statistica.
I observed the scenario described by Carly more than once.
Let me add that we lost a significant Text Mining opportunity because of missing unicode support (where switching settings is simply no option).
And recently I noticed that the Data Function communication between Spotfire and Statistica is restricted to 128 characters (enforces me to decode and encode in Spotfire and Statistica for "special characters").
All this problems could be solved when the *.sta file format would support unicode.
We use the Enterprise version of Statistica and have sites all over the world. We are in the process of adding a Japanese site and have run into this issue with the Japanese characters stored in one of our source systems. Since we use a single server for all sites, changing the Windows language or other settings is not possible. This is preventing that site from being able to accurately pull their data.