L2/09-350 Title: Property Correction for U+A9B3 JAVANESE SIGN CECAK TELU Source: Ken Whistler Date: October 21, 2009 Action: For consideration by the UTC An inconsistent property assignment slipped through the beta review and release for Unicode 5.2. The following character should be removed from the list of characters with the Other_Alphabetic property in PropList.txt: U+A9B3 JAVANESE SIGN CECAK TELU That is the Javanese nukta character, and for consistency with other Brahmi-derived script nuktas, should not be given Other_Alphabetic. This shouldn't make any difference for identifiers, since the character is gc=Mn, and thus is included in all the identifier derivations accordingly. However, it does make a difference for the UCA default collation table, which treats non-alphabetic diacritics (which this is one of) differently. They get a secondary diacritic weight, as opposed to alphabetic combining marks, such as the vowels, which get primaries. I suggest making this change for the next version of the UCD which is released.