[Uim] anthy-6300c is available

[yusuke at cherubim.icw.co.jp]

I am happy to announce newer version of Anthy.
Anthy is text conversion engine to input Japanese text.
(Conversion engines perform inference to determine correct Japanese text)

anthy-6300c seems to have solved anomalies of learning mechanism
in anthy-6300b, so anthy-6300c is the latest stable release.
You can get it from http://sourceforge.jp/projects/anthy/files/

Summary of changes
*Changed search stragey to get higher precision.
 +Use bi-gram(HMM)+Viterbi-algorithm from uni-gram+A-search algorithm.
*Added support of multi-segment word in dictionary.
 +Added 40,000 multi-segment words.
*Reduced memory consumption.
 +Anthy's heap memory consumption is around 100KB (static data) +
  around 150KB when you do actual conversion (depends on length of
  input sentence)

Thank you for users and i18n aware distributors! Your help allow me
to continue the difficult development. (Development of conversion
engine is extremely difficult, it requires developers to use logics,
statistics, heuristics, and black magic. ;) )

P.S.
Thanks also to scim and m17nlib folks to support anthy.
I love scim's motto in scim-im.org `but also a development
platform to make Input Method developer life easier.'