Re: [office] Bottom to top, left to right writing direction

[Regina Henschel <regina.henschel@libreoffice.org>]

Hi Patrick,

Patrick Durusau schrieb am 01-Nov-20 um 16:04:
> Regina,
>
> I'm assuming the following types (from the fifth edition) are at issue:
>
> *****20.1.10.83 ST_TextVerticalType (Vertical Text Types)*****

Yes, it is about mapping to those types.
> **********
>
> Reading the definitions, yes, enabling bt-lr from XSL doesn't solve the
> problem of peculiar definitions in OOXML but it would solve the issue
> for such languages more generally. At least until more guidance is
> issued from the W3C on rendering of CJK languages in texts.

The problem is complex. W3C has specified 
https://www.w3.org/TR/css-writing-modes-3/ about that topic.

> Noting that eaVert and montolianVert are too vague to be meaningfully
> mapped, "where some fonts are displayed as if rotated by 90 degrees
> while some fonts (mostly East Asian)" makes no sense. I have no way to
> know or specify "some fonts."

The above mentioned recommendation has information about it.

> As far as vert top-bottom, left to right and vert270 bottom-top, right
> to left, should we take guidance from 7.29 Writing-mode-related
> Properties, https://www.w3.org/TR/xsl/#writing-mode-related (XSL 1.1)
> and not invent new terminology for it?

It is not enough to add additional values for writing-mode. We would 
need style:glyph-orientation-vertical (20.297 part 3) too. That is 
currently restricted to <style:table-cell-properties> (17.18) and values 
"0" and "auto".

"writing-mode" alone has no means to distinguish between "eaVert" and 
"vert". XSL 1.1 and SVG have "glyph-orientation-vertical" and 
"glyph-orientation-horizontal" in addition
https://www.w3.org/TR/SVG11/text.html#GlyphOrientation
https://www.w3.org/TR/2001/REC-xsl-20011015/slice7.html#glyph-orientation-vertical

> Thinking that if we add writing-style elements sufficient to define
> character styles (perhaps within the standard) that represent common
> character orientations, leaving odd cases to user to define, wouldn't
> that give a target for mapping OOXML vert and vert270?
>
> Realize the present proposal may be inadequate but if the larger issue
> is solving style mappings from OOXML, let's gather those requirements up
> and define them. >

OOXML vert270 is a problem in css-writing-modes-3, because the actually 
needed style:glyph-orientation-vertical="270" has no mapping there and 
style:glyph-orientation-vertical from SVG is considered obsolete by W3C.

Therefore I see these ways:
(A) extend writing-mode with own values which include glyph-orientation, or
(B) extend glyph-orientation in values and usage.

Kind regards
Regina