Last modified: 2011-12-14 17:37:29 UTC
Santhosh Thottingal points out that our plan for language-dependent collation using ICU is not going to work well for certain languages, such as Indian languages. The UCA tailorings are not well maintained. He suggests that we instead use the glibc collation facilities, which are community-maintained and have a much richer set of national collation rules.