Single point side bearing change leads to drastic auto-kerning

[brianbrubaker]

Most of the time, Kern On is spot on with its auto-kerning suggestions for zero kerning pairs, especially since the latest update with the “tighter grip”, and I have found that I’ve indeed needed fewer models to get consistent results.

In general, I’ve also been taking Tim’s wonderful insight about Looking vs. Thinking with KO to heart and adjusting my workflow accordingly (pure gold: viewtopic.php?p=2601#p2601)

However, I’ve recently run into a few instances where I’d adjust the side bearing of a letter and it would drastically change the suggested auto kerning to a point which seems highly unlikely to be a valid suggestion. For example, there was a scenario where if I changed the RSB of the /r by a single point, KO would change from zero-kerning (correct / desired) to -12 (incorrect / undesired).

I realize this may all simply be too contingent on the specific models and designs in a given font. But this was during the process where I was only working with around 20 or 30 models that were completely locked in and double checked by switching each out between Auto and Model to see if KO had any different ideas based on the other models.

Does anyone have any insight into what would lead KO to "think" this would make sense? Is this an issue of insufficient or maybe conflicting models (although this seems unlikely), a bug, or something different altogether?

[Tim Ahrens]

What you describe clearly sounds like a wrong behaviour or Kern On. Would you mind sending me the .glyphs file so I can reproduce it? Thanks!

[brianbrubaker]

Ah, unfortunately I can't find the same file version to reproduce. My backups only go back a month and that just passed literally yesterday from when this exact instance of the file was in use. I went on to find workarounds with different models but hopefully this can still be helpful in some way just to note the behavior.