Re: [css3-writing-modes] inconsistent handling of 'Tr' codepoints in 'text-orientation'

All,

Pardon the intrusion, but it seems that I might be able to offer some insights that may prove to be helpful in this discussion.

In case there was some discussion about U+2329 and U+232A, these code points, in terms of having vertical variants, are now supported by our Kozuka fonts, at least their latest versions. They map to the same CIDs as U+3008 and U+3009, respectively, and thus share the same vertical variants. This is due to canonical equivalence.

Anyway, the real issue seems to be with U+FF1B (full-width semicolon), and to some extent U+FF1A (full-width colon). The conventions for their vertical forms differ by language/region/locale. In Chinese, both characters remain upright in vertical layout, but are shifted up and to the right. In Japanese, only U+FF1A rotates, but U+FF1B remains upright and as-is with no shifting, though Koji once mentioned to me that some people pointed out a use case for a rotated version, hence its Tr property value. About U+3030, the Tr property value seems appropriate.

Regards...

-- Ken

Received on Wednesday, 9 October 2013 11:55:03 UTC