W3C home > Mailing lists > Public > www-style@w3.org > April 2014

Re: [css-text][css3-fonts][proposal] font-side-bearings: normal | trimmed

From: Koji Ishii <kojiishi@gluesoft.co.jp>
Date: Sun, 20 Apr 2014 15:22:34 +0000
To: Peter Moulder <pjrm@mail.internode.on.net>
CC: "www-style@w3.org" <www-style@w3.org>
Message-ID: <5244632B-E27A-456A-B273-B1F8E3372BF7@gluesoft.co.jp>
I think the feature belongs to CSS Fonts rather than CSS Text. CSS Fonts Level 3 is currently in CR, so this will be a request for CSS Fonts Level 4.


On Jan 20, 2014, at 5:48 PM, Peter Moulder <pjrm@mail.internode.on.net> wrote:

> On Sat, Jan 18, 2014 at 10:41:01AM +0100, François REMY wrote:
>> Just for the record, I would like to propose a new property called “font-side-bearings” and which could take the values “normal” and “trimmed”, where “normal” would be the default. The goal of the property is to make sure the first glyph of a line is blackbox-aligned and not moved towards right a bit by the left bearing (in a ltr flow).
>> Context:
>> http://msdn.microsoft.com/en-us/library/windows/apps/windows.ui.xaml.opticalmarginalignment.aspx
>> Thoughts?
> A few random thoughts:
>  - An option for such a feature is to express it in terms of the general
>    effect (usually called margin kerning or optical margins) rather than
>    necessarily require a specific calculation.
>  - The OpenType 'opbd' table (and associated 'lfbd' and 'rtbd' tables) are
>    relevant.
>  - The css-fonts 'font-kerning' property is relevant.
>  - Implementations don't need a property to implement this: font rendering
>    (including kerning) is up to implementations.
>  - The mechanism given in the web page is rather simplistic.
>  - A property might limit the harms of a simplistic approach.
>  - If margin kerning adjustments are to be synthesized (e.g. as described
>    in that web page), then some things to be careful of are preformatted
>    content, spacing characters, and wide combining marks.
>  - Last I looked (a couple of years ago), kerning (including margin
>    kerning) isn't yet well understood, or at least not publicly described.
>    So even if there's a property value that specifies a particular method
>    (as might be useful for predictability among UAs), there should also
>    be an 'auto' value that allows UAs to do better, as the state of the
>    art improves.
> pjrm.
Received on Sunday, 20 April 2014 15:23:08 UTC

This archive was generated by hypermail 2.4.0 : Monday, 23 January 2023 02:14:39 UTC