Support UTF-8 encoding in model annotation/description

[Former user]

The usage of U+2010 ("hyphen") in a model description resulted in U+3F ("question mark") currently occupying that space: Semiphysiological artemisinin PK in healthy subjects. The solution here is of course to use the good old ASCII U+2D ("hyphen-minus"). However, this seems to indicate lack of UTF-8 support which might be a larger problem in the future if it restricts accurate representation of metadata containing non-english characters.