W3C home > Mailing lists > Public > www-style@w3.org > March 2010

Re: [css3-fonts] opentype font feature support

From: Christopher Slye <cslye@adobe.com>
Date: Thu, 11 Mar 2010 11:59:41 -0800
Message-ID: <13EF70A6-C055-492A-8A7D-2CE2D9B1A035@adobe.com>
To: www-style <www-style@w3.org>
I think UAs should respect kerning as any other default "always on" layout feature like liga or locl, rather than making it a UA setting -- although I appreciate the practical reasons why a UA might want to disable it, e.g. below some ppm size. However, this makes me wonder about the language in the Fonts module, section 6, Font rendering properties:

"All of these properties assume that fonts define a given set of default rendering styles ..."

We're talking about defaults dictated by the OpenType features registry, with regard to what is on or off by default, correct? These are not "defined by the font". The font just contains layout features that the UA enables, and some should/must be enabled by default. Perhaps you mean "the font format"?

(Sorry if this has come up already; I've read most, but not all of the thread.)

Can this be made clearer in 6.2, Ligatures? There is no explanation that setting "normal" will have 'liga' enabled and others disabled (which I am assuming is the case). If it's truly dictated on a per-format basis (i.e. if OpenType layout defaults are different than some other format's defaults, and the UAs must follow that) then that should be clearer.

BTW, I assume "normal" is the customary term in CSS, but I do like the word "default" better when it comes to layout features.


On Mar 10, 2010, at 3:30 AM, John Daggett wrote:

> Thinking about this a little more now, I think 'none | normal' makes a
> lot of sense, assuming 'normal' implies "apply metric kerning if data
> available in the font".  If a user agent is supporting things like
> variant glyphs then it's already doing enough processing that
> performing metric kerning is unlikely to have a significant impact on
> performance.
Received on Thursday, 11 March 2010 20:00:14 UTC

This archive was generated by hypermail 2.3.1 : Monday, 2 May 2016 14:38:32 UTC