Sun engineer Xueming Shen talks about why the UTF-8 charset implementation in JDK 7 now rejects non-shortest-form UF-8 byte sequences. Two words: security and performance.