I've seen many distros provide a unicode-data package, containing the human-readable text files from the Unicode Character Database. These are sometimes handy to have around, but I'd be surprised if much software used them directly.