Re: [css3-fonts] @font-face matching and font-style descriptor

Brad Kemper wrote:

> @font-face { 
>  font-family: MyFont;
>  src: local(Arial);
> }
> 
> @font-face { 
>  font-family: MyFont;
>  src: local(Arial);
>  font-style:italic
> }
> 
> @font-face { 
>  font-family: MyFont;
>  src: local(Arial);
>  font-weight:bold;
> }

Right, this with an additional bold italic version will completely
eliminate the possibility of synthetic oblique/bolding.  It would be
somewhat better to organize the use of styles in a design to not use
italic or bold styling explicitly but I realize this isn't feasible in
all situations.

I do think John Hudson's earlier point is an interesting one, do we
want to make control over synthetic styling explicit or is that
overkill?  For example, consider the syntax below: 

  font-weight-synthetic == auto | none

  font-style-synthetic == auto | <angle> | none
  
The initial value for both would be 'auto' and this would correspond
to existing behavior.  The <angle> value would be a number greater
than 0 (and less than ???) representing the obliquing angle to be
used.  Setting font-style-synthetic to <angle> would apply obliquing
when font-style was 'oblique' but not when it was 'italic'.  Setting
either property to 'none' would effectively remove the synthetic faces
from a family.

How to handle families with no italic face is an interesting question.
Existing implementations currently *always* provide a synthetic face
when a font family doesn't have an italic face, so falling back to an
italic face is somewhat problematic.  What happens for Japanese within
an <em> tag, since Japanese font families generally don't include an
italic face?  The reverse case is also interesting to consider -
specifying 'normal' with a font family that contains a single italic
face will always map to that font in all current browser
implementations.

In general, I don't like synthetics because they muddle the
distinction between between real designed typefaces and effects.  But
synthetic bolding and obliquing have been used since the beginning of
web time so we're stuck with them to some degree.

Regards,

John Daggett

Received on Tuesday, 14 September 2010 18:16:02 UTC