The javax.print.DocFlavor class pre-defines types for text data
streams encoded as UTF-16 and UTF-16-LE and equates UTF-16 to UTF-16BE.
This is certainly ambiguous, and probably incorrect.

We should explicitly add the UTF-16BE text cases.
This is a small number of predefined constants in the inner
classes of DocFlavor.java.

Additionally all the constants & strings should be of the form
utf-16be NOT utf-16-be.

This part of the fix needs to be applied to existing utf-16-le entries
which will become utf-16le.