On Tuesday 30 Jun 2009 3:09:15 am Yoshiki Ohshima wrote:
At Mon, 29 Jun 2009 22:30:31 +0300,
Igor Stasenko wrote:
So, do i understood you right: this information (leadingChar) could be useful sometimes. But real world examples showing that its completely useless , at least by now. And often only adds an unnecessary complexity & confusion in handing a text & communicating with world outside the Squeak.
Well, but Squeak in Chinese/Japanese/Korean have been using it and depending on it, so this characterization is not correct.
The essential problem is the encoding of characters (graphemes) which have regional or lingual variations. Devanagari also has such examples. The grapheme 'a' is rendered differently in Bombay and in Calcutta. Should the variations be encoded in the context or in every grapheme codepoint?
leadingChar is just one way to resolve such variations. Personally, I prefer handling such variations as part of the context. A grapheme is a notional element. We need to encode it so that we can input, calculate, combine, compare and sort notional elements. Visual appearance becomes important only for manual edit and print operations.
Subbu