His point that casefolding in a Unicode world is a minefield is completely correct.<p>This classic StackExchange post goes into some of the reasons why: <a href="https://stackoverflow.com/questions/6162484/why-does-modern-perl-avoid-utf-8-by-default" rel="nofollow">https://stackoverflow.com/questions/6162484/why-does-modern-...</a>
I'm not a purist, but working with SQL as a case insensitive but otherwise great and very useful language seems to be really annoying and "shaky", especially when you have interface it with a cade-sensitive interface (e.g. JDBC, pandas).