Positioning non-spacing marks

James Puckett
James Puckett Posts: 2,003
in Technique and Theory
How are the x-direction locations of non-spacing marks determined? I get the impression that these marks are supposed to be positioned to base glyphs with GPOS mark attachments. But it seems like they’re often positioned by locating them far left of zero within their zero-width boxes.

Comments

  • Christoph Koeberlin
    Christoph Koeberlin Posts: 51
    That’s common practice, and I think it makes sense for apps not supporting mark features since the floating accents should follow the letter.
  • James Puckett
    James Puckett Posts: 2,003
    But what logic is used to determine where the accents should even go? Does one just make it kind of work with n and o and hope for the best?
  • Ray Larabie
    Ray Larabie Posts: 1,447
    I figure a misaligned accent over a lowercase o is visually bothersome enough to warrant using it as a measuring stick for non spacing marks. Let's say you have a caron over a lowercase a, e or n that's not quite aligned... it doesn't jump out like a misaligned o accent.
  • Georg Seifert
    Georg Seifert Posts: 677
    Did someone test support for latin mark positioning in different apps?
  • Christoph Koeberlin
    Christoph Koeberlin Posts: 51
    James:
    You need to find a compromise. Making it work with o will be okayish with many other letters, but of course not with extreme widths like dotlessi or w. Nevertheless, a misaligned accent might still be better than none at all.

Categories