[MPEG-OTSPEC] Issues from CSS: Will they be covered by the text shaping WG?

[fantasai]

On 8/12/20 8:46 AM, Peter Constable wrote:
> As yet, a WG with defined scope has not yet been formed. But here’s my 
> inclination:
> 
> 1. Lack of font metrics for writing systems other than Western & CJK:
> 
> yes

This would be great, thanks!

Fwiw, the relevant CSS spec is https://www.w3.org/TR/css-inline-3/

> 2. Italic / Oblique Default Angle
> 
> IMO, the CSS WG doc doesn’t make clear what problem is to be solved. Until the 
> problem is clarified, it should be out of scope for any WG.

The question is, given a font with a variable slant axis, if the author 
requests an "italic" or "oblique" font but does not specify a desired slant, 
how should the font designer communicate its preferred slant for such usage to 
the UA?

> 3. Compatibility problem with 'vert'
> 
> I think so.
> 
> 4. Clarifications for 'vert'
> 
> I think so.
> 
> (Is this really different from 3?)

Yes. The first one is about specific problematic codepoints which may require 
a technical solution. It's unclear how to solve that problem.
The second one is about giving more specific guidance to font designers about 
the orientation of glyphs with 'vert'.

> 5. Clarifications for Bopomofo
> 
> I don’t see a problem here. If tone marks require special positioning, 
> OpenType already provide mechanisms for that. What’s missing?
> 
> If what is missing is documentation to help implementers know what is needed 
> to support a script, then that might be in scope.

If the technical abilities are already in OpenType, then yes, documentation to 
help implementers understand how to support Bopomofo tone mark positioning is 
what's needed (and in the spec particularly, setting the expectation that 
certain features be used for such Bopomofo positioning), and some liaison with 
the Bopomofo font community to help them understand and achieve that.

> 8. Required vs Optional Ligatures
> Same as for 7.

I think the issue here is that the definitions of the ligature features (such 
as 'clig') do not explain clearly whether they can be safely turned off in 
normal typesetting situations (such as applying letter-spacing) without 
corrupting the text.

Basically, from CSS's perspective, features can be
   * mandatory (never turned off because required for correct shaping)
   * optional, on by default
   * optional, off by default
OpenType mentions on vs off by default, but is less clear about mandatory vs 
optional.

~fantasai
W3C CSSWG Invited Expert
editor, CSS Inline Level 3, CSS Text Level 3, CSS Writing Modes Level 3