At 10:35 01/08/13 +0700, Martin_Hosken(_at_)sil(_dot_)org wrote:
PS. One little awkwardness with UTR15 is that the specification (D2) of
composition (C or KC) is that you:
1. Decompose against the current Unicode database
2. Compose against the Unicode database v3.0.0
Well, that definitely should be v3.1.0 now (and is in the
new (since March) version of UTR15).
Notice that you don't compose against the current Unicode database.
It's definitely some kind of awkwardness for implementation, but
it's the right thing to do for stability.
The best way to implement it is to add new decompositions from
the current version of the Unicode database to the composition
exclusion list (which you need to implement anyway).