RE: [css3-text] text-transform values should be grouped

> >> Can we unify ‘large-kana’ with ‘uppercase’?
> >
> > No. Small kana and large kana are actually pronounced differently: in
> > modern Japanese they represent different things. They are in effect
> > different letters that happen to look similar.
> 
> I knew that much – for TU/sokuon at least, for yōon I don’t think it matters
> enough (compare ambiguities in other writing systems). If the difference
> to Western casing is substantial it’s probably better not to have them share
> one property.

As fantasai wrote, the difference is substantial. However, I do not see the needs for the new property. Imagine what "font-variant" properties does. It's related to the font variants, regardless of how it changes. The "text-transform" follows the same rule here. It's related to the text transformation, regardless of how it changes.


> By the way, are normal sized kana really called “large” when contrasted
> with small ones?

Yes. They are usually called just "kana", but when clear distinction between small kana is necessary, they are called "oo-kana", where "oo" is Japanese translation of "large" or "big".


> Also, “kana” being part of the value it would be more
> difficult than necessary to apply this to other scripts if that should become
> necessary.

I understand the point. We should always try hard to unify behaviors in multiple culture where possible. But it's also true that there are a few things that can never be mapped to other concepts. Ruby is another example, and as far as I know, nobody has complained to call it "Ruby", although the name is a Japanese word.

I think we had tried hard enough up until now to unify this into other concepts but my current conclusion is we can't, thus leaving it as is would be the best. You could easily find such examples in the real world; Sushi, Karaoke, or Tempura are probably too hard to unify, aren't they.


> In conclusion, I guess it’s better to let this being handled on the font
> level and allow browsers to fake it.

The use case has two different styling issues: one to make it thicker, and the other to transform it to large-kana.

As you pointed out, the former should be handled on the font level and CSS3 fonts already has the solution for this using font-variant:ruby.

The later one, however, is not a font issue but a code point transformation.

Also I'd like to point out that it is usually printers to make it thicker, and is usually authors or publishers to make it to large-kana, so the decision is independent. This makes sense for CSS3 fonts to handle former, and CSS3 Text to handle later.


Regards,
Koji Ishii

Received on Friday, 8 October 2010 02:30:56 UTC