SQLTeam's favorite Slovenian blogger Mladen (b | t) had an interesting question on Twitter: http://www.twitter.com/MladenPrajdic/status/347057950470307841
I liked Kendal Van Dyke's (b | t) reply: http://twitter.com/SQLDBA/status/347058908801667072
And he was right! This is one of those pretty-useless-but-sounds-interesting propositions that I've based all my presentations on, and most of my blog posts.
If you read all the replies you'll see a lot of good suggestions. I particularly like Aaron Bertrand's (b | t) idea of going into the Unicode character set, since there are over 65,000 characters available. But how to find an illegal character? Detective work?
I'm working on the premise that if SQL Server will...