ISSUE-232: Bottom-to-top text direction

It isn't clear whether this is asking for BIDI support in vertical lines or truly bottom to top vertical lines (as the primary inline progression direction).

If the former, then this is already supported in TTML by using writing-mode='tbrl' or 'tblr' in combination with BIDI content.

If the latter, then this would have an XSL-FO mapping, using the 'bt-rl' and 'bt-lr' writing modes, however, in an HTML/CSS mapping, direct support for bottom-to-top inline progression is not supported; rather, a combination of CSS3 Writing Mode's writing-mode: vertical-rl|vertical-lr would have to be paired with the unicode-bidi: bidi-override and direction: rtl. [I'm not sure this has been tested in using CSS3 WM, but I can check.]

Here's some relevant text from Section 6.3.5 of ST 428-7:

"""
[...]
Bottom-to-top. The Unicode Bidirectional Algorithm shall not be used.
[...]

When bidirectional text is present in vertically rendered text, the Unicode Bidirectional Algorithm shall not be used and the characters shall be rendered in the order in which they appear in the Text element (when reading its content from left to right), in the direction specified by this Direction attribute’s value. When text is vertically rendered, each individual character shall retain its horizontal orientation - the placement of the subsequent character shall be vertically positioned in relation to it.

Note: The underlying thinking behind the above Direction attribute’s values definition is to determine the horizontal rendering direction using the Unicode Bidirectional Algorithm only, without breaking legacy implementations that do not support it at the time this standard was revised. It is expected that horizontal rendering direction would be determined by this algorithm usage only in the future.
"""

428-7 appears to mistakenly assume that the use of vertical text layout concurrently causes text from horizontal writing systems to be set upright rather than sideways (rotated clockwise). However, this is not the default semantics for text-orientation (see [1] and [2]).

Furthermore, 428-7 appears to be in error in proscribing the use of BIDI in vertical text, which *does* apply in the default orientation of Arabic/Hebrew when set in vertical lines (where text is set sideways by default).

In any case, I still am not finding a use case for bottom-to-top writing mode. Except for this mode, CSS will be able to accommodate 428-7, but I fear that a bottom-to-top mode will not be supportable in an HTML/CSS mapping.

[1] http://dev.w3.org/csswg/css-writing-modes/#intro-text-layout
[2] http://dev.w3.org/csswg/css-writing-modes/#text-orientation0

> In any case, I still am not finding a use case for bottom-to-top writing mode.

The latest editor's draft of ST 428-7 shows bottom-to-top direction as deprecated. I suggest we put this issue on hold until ST 428-7 reaches Draft Publication.

WONTFIX. There is no support for bottom-to-top writing mode in CSS, so would not be able to map to HTML/CSS.