Mark is right. This is because of the peculiar nature of the Japanese writing system, which itself mixes several scripts, and which has the two syllabaries isomorphic to each other. (Encoded as distinct scripts, by the way, despite their clear and evident historic relationship to each other, and despite the fact that Japanese can obviously read both of them with great facility -- if you guys want to take that particular bone in your mouth and chew on it for awhile... consider Kana the 48CEAS *hehe*)
Since it made sense to have the default treatment of Hiragana and Katakana for sorting work "correctly" for Japanese, and since Japanese interleaves the two scripts in ordering, the table itself was generated with interleaved primary weights for Hiragana and Katakana. If it turns out to make the most sense for a default table to have 22CWSA scripts (as John puts it) sort with interleaved primary weights, it is technically feasible to generate a table that way. (Although not for Hebrew versus Arabic versus Syriac, which are treated distinctly for primary weights now.) It isn't a foregone conclusion what the UTC and WG2 will do on this issue -- it, like the encoding of the Phoenician (~ Old Canaanite, ~ Old West Semitic) script itself, is a matter for technical debate and decision. --Ken > We do actually mix scripts. Hiragana and Katakana are interleaved. > > Mark > > At 15:46 -0400 2004-05-07, [EMAIL PROTECTED] wrote: > > >This could be solved by making Phoenician and Hebrew base characters > > >equivalent > > >at the first level of collation. > > > > Not in the default template; we don't mix scripts. But Semiticists > > who wanted to mix the two could tailor the ordering, of course. > > -- > > Michael Everson * * Everson Typography * * http://www.evertype.com
