When processing a document with ancient Greek characters, unrtf doesn't handle them properly.

1. It converts them to named entities, like α, instead of their standard unicode character.
2. It doesn't convert all characters. Some I suspect it can't do because there are no named entities for them (like accented characters); others it just seems to skip.

See the attached file, which consists of one word of 7 Greek letters, one with an accent. Every other character is skipped, including a line break.

The solution (for me) would be simply to leave the unicode as unicode.