Re: [css-writing-modes][naming] Inheritance of text-combine

On 07/09/2013 08:22 PM, John Daggett wrote:
> fantasai wrote:
>>>   I'm somewhat confused, the primary use case of tatechuyoko in Japanese
>>>   vertical text is for simple combinations of digits. Why is there a
>>>   need for more complex composition?  Wouldn't an inline block using a
>>>   different writing mode be equivalent?
>> No, it's not equivalent. The need for 'all' is to do anything other
>> than digits. :) Also it's being used currently for EPUB.
> So what's the use case for complex composition?  I'm having trouble
> seeing why we need to worry about complex compositions and inheritance
> given that tatechuyoko spans are typically two or three characters
> long.

Sure, but they are not always digits. :) As Koji has mentioned several
times, there are reasonably common uses of other characters, like letters,
periods, or the number sign.

>>>   I'm not at all clear as to what the "auto" vs. "manual" distinction is
>>>   here.  You're distinguishing the behavior of the 'all' value from the
>>>   'digits' value?
>> Yes. The 'all' value needs to know what element set it, and inheriting
>> it confuses the issue. The 'digits' value needs to inherit to work
>> properly.
> I think it would be simpler to have 'text-combine-horizontal' not
> inherit. I don't think 'digits' needs to inherit unless there's a very
> clear use case where this is needed.

We generally try to avoid having unstyled elements change behavior,
and this is a behavior that makes sense to inherit. For example,
making the day, month, and year into links would disable digits TCY
applied to <date> in this snippet:
which seems unnecessary and surprising.

But I'm OK with this temporarily, I think, if we have to. It's better
than making 'all' inherit, from an implementability standpoint.
However, at some point we may want to make auto-composition inherit,
since IMO that's really the right behavior for this kind of thing.


Received on Wednesday, 10 July 2013 04:14:13 UTC