Re: [css3-fonts] font-feature-settings syntax

Andrew Cunningham wrote:

> Although, I.d be concerned about typing it so tightly to opentype
> and restricting it to ecactly four characters.

Since OpenType is the de facto standard for supporting font features,
that's what's spec'ed out for this version.  That may change in the
future but for now I think it's better to keep this tied to OpenType
since that's what implementations are doing and that's what type
designers are supporting.

> So far I have only heard on one browser experimenting with this
> feature and that browser is supporting both opentype and graphite
> tables.

Firefox ships with support for Graphite shaping pref'ed off by default
(toggle 'gfx.font_rendering.graphite.enabled' to enable).  Graphite
allows anything that you can jam into a 4-byte value and there's no
standardized feature registry.  So it's basically the same as OpenType
without set patterns.  I don't think we can standardize Graphite font
features unless there's some level of consistency across fonts.

> Or is it being proposed that these css properties are opentype only
> and we need to proposed and defined a series of parrallel css
> properties for other font technologies, ie AAT and graphite?
> 
> Esp since the only cross-platform way of implenting fonts for some
> unicode blocks is via graphite in the absences of shapers for
> opentype based font rendering systems. 

The reason it's important to make tags that aren't four-letter ASCII
strings invalid is because this is the mechanism by which CSS assures
some level of future-proofing.  If in the future a new value syntax is
created, we can design it so that it will be invalid in user agents
that don't support it such that authors can use both old and new
syntax until all user agents update to the newest syntax.

This property is the low-level feature syntax to allow direct access
to less commonly used features that may be supported by a given font.
The CSS3 Fonts spec includes subproperties for font-variant that
enable/disable specific features that can easily be mapped onto new
font technologies as they arise in the future.

Cheers,

John Daggett

Received on Friday, 20 April 2012 08:30:52 UTC