Re: [css-writing-modes] text-orientation: sideways-left

Le 2015-01-11 10:26, Koji Ishii a écrit :

> I'd like to propose to:


Here's my opinion on your proposals.

> 1. Defer 'sideways-left'[1] to Level 4, and

I agree. Defer 'sideways-left'[1] to Level 4.

> 2. Make 'sideways' as a value alias to 'sideways-right', not as a
> switch between 'sideways-left' and 'sideways-right' by writing-mode.

I disagree. 'sideways' should not be an alias to 'sideways-right'. Defer 
'sideways' to Level 4 if necessary and if you think 'sideways' would be 
too much to ask/expect from rendering engine manufacturers in short 
period of time, just like Fantasai proposed.

> Currently, WebKit and Blink does not implement this value, and do #2
> above. IE does not implement this property yet. I don't know the
> status of Gecko, but I'd like to sort this out before implementations
> pick different choices for the 'sideways' value.
> Currently the property is prefixed, so we could probably work out if
> truly needed, but I hear no complaints with the current WebKit/Blink
> implementations.
> One more thing to note on this value is that, a concern was raised
> before that all other values of this property only change orientation
> of characters, while this value changes the baseline as well. This
> difference makes implementing this value more complex than others.

Firefox 38.0a1 buildID=20150125230903 implements correctly central 
baseline-alignment when text-orientation is 'mixed' or 'upright' and 
implements correctly alphabetic baseline-alignment when text-orientation 
is 'sideways-right'. So, I am for maintaining this just as it is.


> IIRC the discussion didn't last long, but I think it was not really
> satisfactory for the proposer.
> Other 3 values (mixed, upright, and sideways-right) are important for
> Chinese and Japanese vertical flow, while this value is for rarer use
> cases in other scripts. I propose that this value can be discussed
> further in Level 4, and I wish us not to introduce interoperability
> issues nor slow this spec down because of this value.
> [1]
> /koji

Received on Wednesday, 28 January 2015 03:30:10 UTC