Re: [Devel] Kerning problem with capitals

[Werner LEMBERG]

> >> Advance for "R":         1423 font units
> >> After hinting @12ppem:   8 pixels = 1365 font units
> >> Difference:              58 font units
> >> Corrected "Re" kerning: -92 + 58 = -34 font units (= -0.199 pixels)
> >>
> >> There are probably more intelligent methods, though.
>
> Which of this numbers is rsb_delta? It all iteracts and plays
> together.

No, no, no.  The TrueType interpreter returns *integer* advance width
values only!  Theoretically, it could do something different, but
AFAIK this isn't used.  rsb_delta and lsb_delta are always zero for
the TT hinter.

> > This is an interesting idea.  What about the following
> > reformulation: Get advance width and kerning in font units and add
> > them, then convert it to device units and compare it with the
> > advance width in the device space.  If the rounded difference is
> > non-zero, apply it.
>
> Soon you going to have all the bits and pieces for string rendering
> but no actual string rendering interface.  Oh well, poor freetype
> users will have to read documentation.

:-) See my other mail why I reject such a function today -- as you can
see, I've changed my mind within this half a year.

> Yes, I will help to program this when you agree to break the
> compatibility.

If you really want to help please take the tutorial from
www.freetype.org and add information about lsb_delta and rsb_delta at
the proper places.  You might also add an example function which
implements exactly what you want.  But please also add some warnings
which describe the problems with such a simplistic approach.


    Werner