Well, it depends on what kind of research you do. My former team worked with different text corpora in multiple languages. Proper unicode support is/was very welcome.