W3C home > Mailing lists > Public > www-font@w3.org > April to June 2010

Re: [css3-fonts] low-level font features

From: Chris Lilley <chris@w3.org>
Date: Tue, 22 Jun 2010 12:32:54 +0200
Message-ID: <26795541.20100622123254@w3.org>
To: John Daggett <jdaggett@mozilla.com>
CC: www-font <www-font@w3.org>
On Tuesday, June 22, 2010, 4:48:10 AM, John wrote:

JD> I originally posted this to www-style since it's related to CSS syntax but
JD> I thought I would repost this here to increase the feedback.

I saw it there and had it flagged to reply , but ....


JD> Defining the syntax as a string potentially allows other new font formats
JD> to be supported in the future.  But it is slightly odd to be defining it as
JD> a string but then specifying a syntax for OpenType fonts only.

Yes, it is. There should be some extensibility, although it should also be clear that OpenType is the primary focus.

JD> To simplify this a bit, I'm thinking of revising this to:

JD>   font-feature-settings:  <featuretaglist> | <string>
JD>   <featuretaglist> = ident[( <integer> )]? [, ident[( <integer> )]?]*
JD>   
JD> For OpenType fonts, <featuretaglist> must be used while for other font
JD> formats either <featuretaglist> or <string> can be used, without no
JD> defined format for non-OpenType fonts.  The <integer> value must be 0 or
JD> greater; if omitted the value of 1 is assumed.  Feature tags that don't
JD> exist are simply ignored.

That looks reasonable to me.

JD> An example using the current syntax:

JD>   /* Turn capital spacing on, turn proportional kana off */
JD>   font-feature-settings: "cpsp=1, pkna=0";
JD>   
JD> Using the proposed revised syntax:

JD>   font-feature-settings: cpsp, pkna(0);

I like the proposed new syntax better than the old one. It seems more natural, and the functional notation mirrors similar usage in other parts of CSS.

JD> For OpenType fonts, the example below would be valid syntax but would be
JD> effectively ignored since neither of the tags are valid OpenType feature
JD> tags:

JD>   font-feature-settings: this(5), that(0);

That seems ok (and allows deploying stylesheets in the future that deal with multiple revisions of OpenType.

JD> This syntax feels more CSS-like and eliminates the use of quoted strings
JD> for OpenType fonts.  I don't know if it can completely express feature
JD> settings for other font formats. 

I suggest concentrating on a couple of likely-to-be-used other formats rather than worrying that some future format might be incompatible. if that happens the spec can always be revised if necessary.

JD>  The ability to specify a <string>
JD> allows some room for defining a way of specifying feature settings for
JD> future font formats.

JD> It would be very helpful to hear whether better support for AAT font
JD> features should also be defined here (especially folks from Apple) or
JD> if there are other font formats that should also be considered.

I agree that AAT and Graphite, both applied to sfnt fonts, would be two real-world use cases to consider here in terms of applicability of the syntax.





-- 
 Chris Lilley                    mailto:chris@w3.org
 Technical Director, Interaction Domain
 W3C Graphics Activity Lead
 Co-Chair, W3C Hypertext CG
Received on Tuesday, 22 June 2010 10:32:57 UTC

This archive was generated by hypermail 2.4.0 : Friday, 17 January 2020 22:37:34 UTC