Kern On

(Apologies if this has been discussed before)

The ‘Generating class kerning’ phase takes a long time and is only running one (?) thread. Can this not be sped up?

I spent considerable time improving the speed of Kern On but there are limits as to what you can do. To me, the crucial aspect was the speed of the tool while you are working on your kerning, the reaction time as you set or change models, for example. I believe this is fairly snappy now.

During finalization, the general autokerning is done with multi-threading and I am somewhat proud of this achievement. Please be kind and don’t complain about the class kerning generation _not_ being multi-threaded. It can’t be separated into parallel tasks as everything is connected in that phase. At least, the class kerning generation is running in a background thread (which is also not entirely trivial) so you can continue using other apps on the computer while you wait for Kern On to finish.