Re: [CSS3 Text Layout] working draft

Thanks for your comments.

Paul Nelson (ATC) wrote:
> Christopher,
>
>   
I'm not familiar with Michel Suignard's documents, could you provide a 
reference?
> Thanks for your mail. From my perspective, that which is already defined
> for XSL-FO and what was previously defined by Michel Suignard's
> documents are able to produce correct results.
>
>   
The people who you interacted with might have found it "weird" for 
reasons other than simply being vertical, such as poor layout engines or 
poor content design due to lack of experience with vertical web layout.  
In instances where I have encountered vertical text I didn't find it 
"weird" after a fairly short period of adjustment.

Now that I think about it though, the block progression direction should 
be defined by the flow direction of the Box Model so block-progression 
is not needed as a property of Text Layout at all.
> 1. I would classify neither Korean or Chinese as being "left". Also, my
> interactions with the Japanese user community is that they consider it
> "weird" to have vertical text on the computer...even though nearly every
> book they purchase is laid out vertically. Thus, the categorical
> classification of block progression for text can only be as an example. 
>
>   
Using a single attribute for horizontal and vertical inline directions 
is compact, but the coupling between horizontal directions and vertical 
directions doesn't make sense.  There is no reason for right to left 
scripts to be laid out bottom to top by default (vertical direction 
corresponding to the right to left horizontal direction) when all other 
scripts are vertically laid out top to bottom by default.

The second figure on page 6 of the Robust Vertical Text Layout paper 
(http://www.unicode.org/notes/tn22/) is an example of text which is 
rotated counterclockwise in a vertical layout.
> 2. The HTML and CSS specifications already have properties (dir or
> direction) to take care of left to right or right to left flow on the
> line. The vertical line is still either left to right or right to left.
> Only the escapement of the line has been changed. The confusion is the
> rotation of most East Asian glyphs when the line is vertical.
>
> 3. All lines of text rotate 90 degrees clockwise when text is laid out
> vertically. In fact, Mongolian should be considered a "left to right"
> language as it is laid out left to right when it is horizontal and is
> also laid out from top to bottom when vertical. Arabic is bottom to top
> when laid out vertically. (Are you confused yet?)
>
>   
Decoupling the block progression direction and inline progression 
direction allows changing the direction of one without having to keep 
track of the other.  For example switching vertically laid out Latin 
from top-to-bottom to bottom-to-top without having to know whether it is 
embedded in Mongolian or Chinese/Japanese/Korean.

The inline direction of the writing-mode property seems redundant with 
the dir property you mentioned earlier.
> 4. Currently, the writing-mode property (lr-tb, rl-tb, tb-lr, tb-rl,
> etc.) take care of the line flow and line stacking in a coherent manner.
> The writing-mode property has been supported since IE5 (except for tb-lr
> which will be added in a future version).
>
> 5. The glyph-orientation property of XSL-FO is one property that should
> be added to CSS at some point to allow for authors to specify the
> orientation of the glyphs on the line so Latin can be stacked
> vertically, etc.
>
>
> Paul Nelson
> (Beijing)
>

Received on Thursday, 4 January 2007 10:28:56 UTC