Online recognition of handwritten characters is regarded as one of the
key technologies in the multimedia era. The system which we developed
recognizes handwritten gestures for editorial operations as well as c
haracters and is now used in electronic organizers and word-processors
. Simultaneous selection and positioning of editorial operations with
a pen effects faster text edition. The recognition rates are 98.6% for
the set of about 3600 Kanji characters including all of JIS level 1 a
nd a part of level 2 characters, 96.0% for the set of about 400 non-Ka
nji characters and symbols, and 99.5% for 15 patterns of gesture.