Differing kerning for numr and dnom

Post Reply
User avatar
SCarewe
Posts: 102
Joined: 23 Apr 2021

Differing kerning for numr and dnom

Post by SCarewe »

Hello, I am running into an odd issue. I have my numr glyphs set up as components (align=3), so exact same sidebearings as my dnom glyphs. Kern On wants to kern my numr differently (slightly tighter) than my dnom. Why? I have them both in the same special spacing group. Saving and re-opening the file didn't change anything, just like trying a different special spacing group. Do I need to re-define models for every set of glyphs (dnom, numr, superior, inferior)? I thought this was intended to be taken care of. Thanks for any pointers!
Attachments
Screenshot 2023-03-16 at 15.55.22.png
Screenshot 2023-03-16 at 15.55.22.png (51.32 KiB) Viewed 2659 times
Screenshot 2023-03-16 at 15.55.03.png
Screenshot 2023-03-16 at 15.55.03.png (60.12 KiB) Viewed 2659 times
User avatar
Tim Ahrens
Site Admin
Posts: 407
Joined: 11 Jul 2019

Re: Differing kerning for numr and dnom

Post by Tim Ahrens »

Thanks for pointing this out. It seems you have numr and dnom letters, right? Never knew that was a thing. ;-) The reason for the inconsistent behaviour is that Glyphs says that these are lowercase letters. That’s somewhat debatable, I’d say but never mind. As (real) lowercase gets a special treatment in that the zones above the x-height and below the baseline are de-emphasized, this led to inconsistent behaviour between these sets of “lowercase” letters (numr and dnom may have different portions in these zones).

The good news is: I just re-worked the handling of dnom and numr in Kern On, and the “lowercase letter” attribute is stripped internally so these glyphs don’t get the de-ephasis. Also, those glyphs triggered by dnom and numr only get kerning with themselves, plus numr gets kerned to all partners before ½ in the list of Unicode pairs. Similarly, all dnom get kerned towards all partners after ½. Plus, now, all numr and dnom are kerned to the fraction slash (before and after, respectively).

You can get the update from https://kern-on.com/update/

Hope this fixes all problems. If you experience any further issues please let me know!
User avatar
Tim Ahrens
Site Admin
Posts: 407
Joined: 11 Jul 2019

Re: Differing kerning for numr and dnom

Post by Tim Ahrens »

Btw, I believe this makes one of your scripts unnecessary:
https://github.com/eweracs/glyphs-scrip ... topairs.py
User avatar
SCarewe
Posts: 102
Joined: 23 Apr 2021

Re: Differing kerning for numr and dnom

Post by SCarewe »

Yes, this font has numr and dnom letters (don't ask me why, the only use-case I was able to discern was writing Vème siècle in French, with ème being in numerators). Anyway, with the update, this works as expected now. Same for fraction kerning. Thanks a lot! I will very happily finally retire that script.

On that note: I strongly suggest adding numr letters to dashes as autopairs.
Example: VIe–VIIe siècle (with the e being a numr in each case).
Attachments
Screenshot 2023-03-27 at 23.33.05.png
Screenshot 2023-03-27 at 23.33.05.png (29.71 KiB) Viewed 2627 times
Post Reply