RE: Shaping characters in upright orientation in vertical text flow

Thank you everyone for very prompt replies.

After looking at your replies, I started to wonder, maybe the way I asked the question wasn't very good, so please allow me to rephrase.

We editors expect most non-East Asian authors would use "text-orientation: sideways-right" as explained in the figure[1]. In this style, all non-East Asian characters such as Latin or bi-di renders sideways, rotated by 90 degrees, as you all recommended. So, that's good to have this many confirmation.

We have another option here, "upright," that sets Latin letters upright. This is a value that are not often used in Latin either, but we have this value because there are some use cases such as signboards.

The question is about how we should render shaping characters in this style. If you never want shaping characters appear in upright, and therefore ignore "upright" style and prefer fallback to "sideways-right", that's doable, but then you'll lose the capability of setting them upright ever. If rendering such way is totally incorrect and you want to prohibit authors to use such an option, I think that's a good design.

If, even rarely, some authors may want to render shaping characters such as Arabic in upright, I think it's a nice option to keep as is.

So, from what I read, I understand what everyone says are:

    "sideways-right" is the most common way to set shaping characters
    in vertical text flow, but "upright" is sometimes used for shaping
    characters such as Arabic, and when doing so, using isolated forms
    is correct.

Do I understand your answers correctly?

Thanks for your support again, as always.



From: CE Whitehead [] 
Sent: Tuesday, January 17, 2012 9:56 PM
To:; Koji Ishii
Subject: RE: Shaping characters in upright orientation in vertical text flow

Date: Tue, 17 Jan 2012 10:25:12 +0200
Subject: Re: Shaping characters in upright orientation in vertical text flow

> I am not an Arabic expert, but what I seem to remember is as follows: 

> a) Arabic letters oriented upright should be in isolated shape 

> b) This is a very unseemly sight for an Arabic reader and authors should avoid this mode of presentation. Arabic letters rotated 90
> degrees counterclockwise so that the text is read from top to bottom while the letters can be connected is a much better option.

I would tend to agree with Mati that Arabic words/text are/is not for the most part presented with the characters arranged vertically in isolation; the alphabet can be presented this way of course-- I've seeen the latter. 


--C. E. Whitehead 
 > Shalom (Regards),  Mati
 >     Bidi Architect
 >     Globalization Center Of Competency - Bidirectional Scripts
 >     IBM Israel
 >     Mobile: +972 52 2554160

From:        Koji Ishii <> 
To:        "" <> 
Cc:        "'WWW International' (" <>, John Daggett <> 
Date:        17/01/2012 06:26 
Subject:        Shaping characters in upright orientation in vertical text flow 

Hello, I've got an item that I need your help.

CSS Writing Modes Level 3 has the "text-orientation" property[1]. With that, you can set glyph orientation in vertical text flow.

The issue is about how to render shaping characters when "text-orientation: upright" is applied. Please scroll down the spec a little bit to see "Figure 12. 'text-orientation' values"; "upright" is the one I'm talking about. You see all characters including Latin are upright in this style.

How do you expect shaping characters look in this case?

Currently, the spec states "Shaping characters from such scripts are shaped in their isolated forms." This is primarily from fantasai's investigations.

Another source to support this behavior is how Excel renders its vertical text flow[2].

There're other options such as "always keep them sideways (i.e., rotated by 90 degrees.)" This is the behavior usually done by using "text-orientation: sideways-right"[1] (see Figure 12 for examples,) but I understand there're some scripts that can never be written in upright and therefore renders the same way as "sideways-right" even when author applied "upright".

I personally have no idea which one is the right behavior here.

John Daggett in his recent mail pointed out that[3]:
> For 'upright' the spec currently states "Shaping characters from
> such scripts are shaped in their isolated forms." This means that
> 'upright' applied to Arabic in vertical text would break the shaping.
> I'm really not sure that this is the right behavior, I think this behavior
> is why Microsoft was talking about making an alternate proposal for
> UTR50.

Could anyone give us your opinions on this? Thank you for your support in advance.



Received on Tuesday, 17 January 2012 13:55:24 UTC