W3C home > Mailing lists > Public > www-style@w3.org > May 2017

Re: [css-writing-modes] Re: Fwd: Is it possible to choose rotational direction of vertical script if I want to force them to display horizontally?

From: fantasai <fantasai.lists@inkedblade.net>
Date: Tue, 2 May 2017 15:49:18 -0400
To: www-style@w3.org
Message-ID: <45dfcd21-775f-5acd-5e2a-dccba09d2bff@inkedblade.net>
On 03/11/2016 06:17 PM, gfb hjjhjh wrote:
> Then is it possible to specify something in CSS writing mode level 3 to change the line orientation?

Only vertical-lr vs sideways-lr. There is no ability to change
the line orientation in horizontal text.

> The line orientation example in https://www.w3.org/TR/css3-writing-modes/ section 1 suggest that the right hand side is the up
> direction, however according to what I've seen, it's true for the case when it's arranged together with other horizontal-tb,
> ltr text, but I'm not sure about what'd happen if those Mongolian text are placed within Arabic or Uyghur text, and I think
> those who can write Mongolian text and are not influenced by default direction in computer, nor modern ltr publication (which
> are all ltr by default) might not consider right hand side as top.
>
> The draft also stated that, for Mongolian text, "In horizontal text, they are typeset in a 90° counter-clockwise rotation from
> this orientation." which I am aware of it is the genral approach to handle Mongolian text in horizontal emvironment nowadays
> but I don't know if it is 100% universal.
>
> I think adding an attribute for these nonrotatable vertical text so that user can choose if it want the left side or right
> side to be top/upright when it's forced horizontal or when there's elements like overline and underline might be a good idea?
> That's like specifying if you want your English text to appear with the character "A"'s head pointing left or right when it
> have to be rotated.

The case of clockwise-rotated Mongolian seems rather theoretical
at this point, and we don't add features for theoretical or archaic
use cases to CSS. It is not worth the implementation complexity.
If at some point this becomes a significant issue, we'll solve it
at that point.

> A bit more note: In the CR's Appendix A it said Old Turkic is ttb
> in its Vertical direction but there's some other pages who say it
> is btt. Which one is correct?

I'm not a scholar of this language, but the evidence seems to point
to ttb-lr. A comparison of the Unicode code charts
   http://www.unicode.org/charts/PDF/U10C00.pdf
and this inscription
   https://upload.wikimedia.org/wikipedia/commons/d/d6/Gok_turk_Epigraph_Copy_in_Gazi_University_Ankara.jpg
seems to indicate that it is written top to bottom, left to right.
(Note that Old Turkic is encoded as RTL, and the glyphs are rotated
counter-clockwise.) This is also backed up by the description in the
Unicode Standard:
   http://www.unicode.org/versions/Unicode9.0.0/ch14.pdf
Additionally, we can see on the inscription that the bottom edge
is ragged and consistently ends in punctuation, which is a good
indication that it is the end, not the beginning, of the lines.

> And actually in the current status what is the expected behaviour
> for Mongolian text in vertical-tb, rtl environment? would they
>form in a wrong way or force itself ltr?

There is no such thing as vertical-tb. In an rtl environment, the
Mongolian text will remain ltr, just like English and Chinese remain
ltr. Punctuation before and after the Mongolian text run, being
bidi-neutral, may become scrambled, however. You can look up the
Unicode Bidirectional Algorithm for more information on exactly how
this works.

~fantasai
Received on Tuesday, 2 May 2017 19:49:54 UTC

This archive was generated by hypermail 2.3.1 : Tuesday, 2 May 2017 19:49:55 UTC