I'm late posting, but case-sensitivity is just one of many ways that room nicks can be confusing. "StPeter" vs "St.Peter" vs "St. Peter" vs "St Peter", or even "StPeter " -- and then there are all the visually confusable Unicode glyphs. ("StΡeter".) It's not obvious to me that case-folding is the worst of these.
That's true. That's why Textshell's proposal of leaving it up to the server to decide what's duplicate and what not is the best way to go. A simple implementation does no checking at all, whereas a very complex implementation could use an algorithm to calculate similarness of nicknames (taking into account the whole Unicode set and spaces/dots/characters in between characters, ...), and reject the ones that are too similar. cheers, Remko