This talk discusses some problems one encounters in manipulating multilingual Unicode text. The problems include using multiple fonts to display Unicode plain text, dealing with neutral characters, deunifying characters that look different in different scripts, dealing with complex scripts, using keyboards to enter Unicode characters conveniently, keeping compatible with previous character sets, glyph variants, and navigating through text that includes "multicharacters" such as Unicode-surrogate pairs, combining-mark sequences, and variable length end-of-paragraph marks. Solutions to some of the problems will be demonstrated using the RichEdit 3.0 controls shipped with Office 2000 and to be shipped with Windows NT 5.0. Glyph variants and surrogate pairs comprise a couple of ongoing topics that lend themselves to several possible approaches.