Started a new thread on the Kerning topic here as the old subject line 
was inappropriate.

On Tue, 6 Dec 2005 12:15 am, Luca Furini wrote:
> Manuel Mall wrote:
> > I wonder if the same argument does apply to kerning as well? The
> > moment you change font-size, text-decoration, background-color,
> > alignment and the like, which is what fo:inline is mainly for, you
> > create a barrier with respect kerning. Or to put it differently, I
> > believe kerning applies only to "like" characters, same as glyph
> > merging.
> Not sure here: if we want to use kerning between, for example, "VA" I
> think it would be even more desirable to use it if the V has a bigger
> font size, or the A a smaller one; it could maybe need some heuristic
> to handle a kerning between character with different sizes, but I
> think it would be a desirable feature.

There seem to be different kerning algorithms around. The simplistic 
approach seems to be to use the kerning information stored in the font. 
That information can only be applied to characters with the same font 
size. And yes it does suffer from the problem you mentioned above. 
Typesetting programs may provide additional kerning algorithms (apart 
from allow the user to manually adjusting kerning). For example Adobe 
InDesign has something like "optical kerning" where the kerning 
information is calculated based on the shape of adjacent characters. I 
would think something like that would be outside the scope of FOP. In 
the end XSL-FO has the letter-spacing property which users (and 
programs generating XSL-FO) can use to adjust kerning.

> Nonetheless, I agree that the think may become strange with different
> backgrounds or vertical alignments ... maybe we should define which
> properties "break" a kerning pair ...
IMO FOP should limit itself to:
a) Use kerning only for consecutive characters within the same fo
b) Limit itself to the kerning information in the font
c) Only apply kerning if the letter-spacing property has the value 
"normal" (and the font supports it)

> Regards
>     Luca


Reply via email to