…seems to be a thing that might be possible, no? I can see that A and V might have equivalent sidebearings but very different kern values with (say) T, but in many cases they would be identical: O, Q C D etc. Might a simplified/streamlined “kerning/sidebearing the same” combo be a possibility?
This is where kerning groups come into play. If you are judicious about setting up kerning groups, this works out. If you establish a side bearing for O and assign the value for LSB for C as =O and put them in the same kerning group, follow like pairings, and it works out.
I do this. What I’m suggesting is that there is ONE setting for both sidebearing and kerning – so that setting a sidebearing equivalence ALSO sets a kerning equivalence. This would hold more often then not, I’m guessing
The side bearings of A can be the same as for V but if that triggered the same kern values, you would have collisions on pairs like VV
A lot cases could be handled with the same setting. But you need to be able to set different values. Having a system that can switch between the two cases would complicate it much more.
And each field has different requirements/possibilities. (Metrics keys can calculate and add measurements heights.
You are right, Georg, you’d have to have some kind of exceptions/override, which might just make it more complex. Just a thought, as in many cases they are the same so it seemed like a possible simplification of the process (which is always welcome!)
There’s a relatively simple solution that could be implemented: the user could just link the kern values of any selected group/glyph pairs: @A @O = @V O = O V, etc. No further complications and exceptions, because the user defines what exactly is linked to what. 