W3C home > Mailing lists > Public > www-style@w3.org > December 2016

Re: [css-text-decor] Emphasis marks and font variant

From: fantasai <fantasai.lists@inkedblade.net>
Date: Fri, 23 Dec 2016 21:34:58 +0330
To: Ken Lunde <lunde@adobe.com>
Cc: Xidorn Quan <quanxunzhen@gmail.com>, www-style list <www-style@w3.org>, John Daggett <jdaggett@gmail.com>
Message-ID: <5caf30e1-52f5-fb10-6242-3936f1d8be16@inkedblade.net>
On 12/23/2016 12:57 AM, Ken Lunde wrote:
> Elika,
>
> Two things should happen when applying the 'ruby' GSUB feature. One is to get the substituted glyphs, which are specifically tuned for ruby use. For some fonts, such as Adobe-Japan1-4 and greater Japanese ones, the 'ruby' GSUB feature will return different glyphs. For other fonts, such as Kenten Generic, it's a no-op because the glyphs are already tuned for ruby use. The other thing that needs to happen is for the client to scale the glyphs, usually to 50%. This is outside the scope of the OpenType 'ruby' GSUB feature.
>
> Anyway, to answer your question, I think that the answer is yes. For some fonts, different glyphs will be returned (Adobe-Japan1-4 and greater), and for other fonts, the feature is not supported, which results in a no-op (Kenten Generic). In both cases, the glyphs, whether substituted or not, need to be scaled.

Thanks for the info! I've updated to apply 'ruby' as you recommend,
and will run it by the WG for the next publication. The spec now reads

# The marks should be drawn using the elementís font settings with
# the addition of the ruby feature and the size scaled down 50%.
# However, since not all fonts have all these glyphs, and some fonts
# use inappropriate sizes for emphasis marks in these code points,
# the UA may opt to use a font known to be good for emphasis marks,
# or the marks may instead be synthesized by the UA. Marks must
# remain upright in vertical typographic modes: like CJK characters,
# they do not rotate to match the writing mode.

~fantasai
Received on Friday, 23 December 2016 18:05:29 UTC

This archive was generated by hypermail 2.3.1 : Friday, 23 December 2016 18:05:30 UTC