It would nice if we could use the full set of UTF-8 characters for problem description and sample data. Currently, playing with fontencodings can get some non-ASCII characters rendered properly, but it is tedious to use and is problematic with sample data.
To fix this, we probably need to change the LaTeX engine that we use, but it would greatly simplify the use of different characters. It's the 21st century, lets go 🚀