ICU stores most UnicodeData.txt properties in its uprops.dat, currently some 23kB (Unicode 3.0).
This does not include character names, which are in unames.dat, currently some 83kB.
There is currently a bug about wrong properties for the last 1k chars in plane 15 & 16 (I will try to fix this before ICU 1.8), but otherwise it works fine for all of Unicode.
It's open source.
http://oss.software.ibm.com/developerworks/opensource/icu/ubrowse?k=100000
markus
Peter_Constable@sil.org wrote:
> Does anyone have compact implementations that are open-source (or otherwise
> share-able)?
This archive was generated by hypermail 2.1.2 : Tue Jul 10 2001 - 17:21:18 EDT