Re: About Chinese Font Generic Family

In Chinese print books, Kaiti（楷體）is more generally used than Heiti（黑體=gothic in Japanese）.

Do you mean for emphasis purposes, when body text is Ming? So what you want is to distinguish Kaiti from Ming?

If css font module level 3 count Kaiti as Serif font, authors should list every usable Kaiti in @font-face.

What I talked in the conf call today was that, in Japan, authors tend to care more than generic, so they actually prefer listing 5-10 fonts in @font-face. This could be sometimes multiplied because UA’s font name recognition issue, which should be solved when CSS Fonts Level 3 is implemented.

In EPUB, authors prefer generic families (or e-book stores recommends,) but none of e-readers have other fonts than one Mincho and one Gothic, so I don’t hear voices for more generic families.

So my proposal is: let Kaiti in Chinese as cursive, that may solve practical problem.
I think put Kaiti together with Gyosho（行書）and Sousho（草書）

fantasai supported your proposal and John said he’s open[1], so I think it’s just a matter of nobody followed up and lost. You could follow up to the thread. John was busy on implementations, but I believe he’s back to the WG now.

More information about Webfont. A woff file contains 39,000 characters up to 11MB.
There’s some services with dynamic subsetting in Taiwan, but not generally adopted.[2,3]

You could try getting statistics of unique characters in a web page (or are you talking about EPUB?) Japanese may need less unique characters thanks to Kana, but I don’t think most single document do not require 39,000 unique characters. If 390, a simple math would give you a 110KB, which is quite possible size to download. If 780, it’s still 220KB.

If subsetting is useful for Chinese too, it’s probably better to talk to providers. I think it’s useful, since you might have done it already in PDF.

[1] http://lists.w3.org/Archives/Public/www-style/2013Nov/0274.html


/koji