On Tue, 23 Aug 2011 10:02:05 +0800 li bo <libo....@gmail.com> wrote: > ...But I don't know why user must take > a paragraph as a unit to determine the embedding levels. Why can't i > shape the text first and then wrapping the line, and determining the > embedding levels for characters within a line. finally, reordering > the characters within a line. If a paragraph is too long, i think > it's a big memory occupied. This would be a limit in embedding > system such as mobile phone.
There is one problem with processing the text line-by-line - W5 'ET ET EN -> EN EN EN' and N1 can both require one to examine the last character in the paragraph, however many lines away it be. Apart from that, the process is just simpler to describe if one performs the algorithm step by step. Richard.
