-----BEGIN PGP SIGNED MESSAGE----- Hash: SHA1 Hi Stefan and others,
>> 1. Horizontal, that is, left-to-right (LTR) versus right-to-left >> (RTL). >> 2. Vertical, that is, top-to-bottom (TTB) versus bottom-to-top (BTT). >> 3. Priority of direction (e.g. (LTR, TTB) versus (TTB, LTR)). >> [...] An elaboration of the directional override mechanism to >> handle vertical directionality would have to take priority into >> account as well. Instead of two directionalities, LTR and RTL, the >> Unicode Standard would have to consider eight. This is not quite sufficient, I'm afraid. An interesting approach is, for example, taken in the Omega typesetting system, an extension of TeX, where they deal with a total of 32 writing directions by describing writing directions by three variables: 1. "Top" side of the page (one out of four) 2. "Left" side of the page (one out of two, since it *does* have to be orthogonal to 1) 3. "Top" of character (one out of four). This is needed if you want to embed, say, RTL and LTR text in the middle of a TTB paragraph, such as embedding other scripts in Mongolian, or rotated plain script in vertical CJK. This gives you 32 combinations of what is written into which directions and how glyphs have to be aligned. It's described in a paper on http://omega.cse.unsw.edu.au:8080/papers/directions.pdf, but be kind to the poor server :-) The most "common" direction combinations given there are the following eight: TLT - Left-right scripts, horizontal CJK TRT - Right-left scripts RTT - Vertical CJK, upright left-right scripts in vertical CJK RTL - Mongolian in vertical CJK RTR - Rotated left-right scripts in vertical CJK LTL - Mongolian [This actually depents on the orientation of Mongolian glyphs] LTR - Rotated left-right scripts in Mongolian LTT - Vertical CJK in Mongolian Even this 32-direction system does not cover cases such as Mayan where scripts follow the rebus principle, where text gets embedded inside other glyphs or where text is written "boustrophedon" ("as the ox plows", i.e. line by line in both directions). Actually, this is a DTP issue. It has very little to do with the information content of the text, instead it's rather a presentation thing, which is why IMHO it does not belong into Unicode, but that's a matter of discussion. SP> As an alternative solution, the current switches could be considered LTR, SP> TTB and RTL, TTB. Then 6 other code points would be necessary for the other SP> directions. As said, this is not quite enough. Admittedly, some directions are rather arcane, but believe me, it will be possible to dig up some ancient document or other which is written into any odd combination of directions. If Unicode wants to achieve complete representation of any kind of presentation of text direction here, it's going to be a pretty rough job. Kind regards and happy new year to everyone (at least everyone following the Gregorian calendar, that is :-) Philipp mailto:[EMAIL PROTECTED] __________________________ Out of memory / We wish to hold the whole sky / But we never will -----BEGIN PGP SIGNATURE----- Version: GnuPG v1.0.6 (MingW32) Comment: Freedom of the press is limited to those who own one. iD8DBQE8Lcd/AFQhKhQ6O0kRAp8tAJ9lfrU19gkm4de/b9wO/ucRjj18GwCgkRgx B3G6wlPGAmbF44Y3imrZGc0= =ZBG9 -----END PGP SIGNATURE-----