From jvoisin (CC BY-SA)
Some artists will benchmark your utf8 support [and beyond]:
[…] Magma and their Kobaïan made-up language featuring some custom diacritics that don’t exist in utf8.
Of course, see also the famous Falsehoods Programmers Believe About Names.
This also reminds me that on BC BY-NC-SA I’ve noticed how bad Bandcamps slugify is: 無限大 becomes –14 (the 14 is just an anti-hash-collision counter, I believe)