Re: Memo from ruby disucssion with Roland

fantasai, Thu, 23 Feb 2012 03:29:43 +0100:
> On 02/23/2012 03:02 AM, Roland Steiner wrote:
>> On Thu, Feb 23, 2012 at 07:24, Leif Halvard Silli:
>> 
>>> 2. Do we agree that column-major - what is in HTML5 now - should be
>>>    non-conforming?
>> 
>> I don't think that's an option as it would break existing pages. A solution
>> should have graceful fallback to both the current standard as well as to no
>> <ruby> support.
>
> Agreed. 

Unless *I* misunderstood, then Roland misunderstood what 'column-major' 
meant. Hence, I am also uncertain what you mean - see my reply to him.

> HTML5 ruby, as it stands now, can be folded into a either a
> row-major model or a column-major model,

Perhaps only my own problem, but I don't understand how one can — and 
what it means — to 'fold' HTML5's current model - column-major,

   <ruby><rb/><rt/><rb/>
         <rt/><rb/><rt/></ruby>

into row-major:

   <ruby><rb/><rb/><rb/> 
         <rt/><rt/><rt/><ruby>

If we think that the row-major model is the way to go, then it is the 
one that should be conforming. What are the benefits of keeping the 
column-major model conforming? I see only one: Not disturbing authors. 
Are there others?

The negative sides of column-major has to do with accessibility - in 
the widest sense of the word. 

Also, we think tactically — at least if Ian and/or the HTMLwg chairs 
are to be convinced, then to say 'but you can also use column-major', 
will only seem as if we have no good reason for saying that row-major 
should be the way forward. Also, it is often said that it is simplest 
if there is only one way to do it. And this seems like such a case.

> so aside from the parsing
> algorithm (Bug 13113), it does not create a conflict with moving
> forward on a row-major model. It just means that in a row-major model,
> the HTML5 markup is only capable of producing group ruby, not jukugo
> ruby. But a row-major model can easily be a superset of what's in
> HTML5 today.

I don't understand how to see row-major as a superset of column-major. 
Please note that HTML5 does not limit itself to <ruby>rb<rt></ruby> — 
HTML5 currently also allows <ruby>rb<rt/>rb<rt/></ruby>.

> http://www.w3.org/Bugs/Public/show_bug.cgi?id=13113 (in which
> hsivonen asks for parsing algorithm changes that would make HTML5
> future-compatible with row-major ruby markup)

When you refero to that bug, then I think that you consider row-major 
ruby as a ruby which has support for <rtc> - and thus double sided 
ruby. However, that is not what Koji meant by row-major. It would be 
nice if we could decide about the model for single-sided, before 
deciding <rtc>.
-- 
Leif H Silli

Received on Thursday, 23 February 2012 03:01:53 UTC