They happen when lines meet where the width of the feature isn’t an exact multiple of the extruder’s width and the printer has trouble filling in the void. Usually they’re fairly invisible, but when printing white on black, they stand out like a sore thumb.
I’m wondering if there’s a good simple way of avoid this problem in the slicer. The ultimate fix of course is to print a square sheet of white PLA under the white letter, but I’d rather not mess with my model because it’s quite complex already.
Write your gcode by hand and you can do curves thus ensuring lines won’t meet. I last wrote gcode by hand in the 1990s, and I’m pretty sure industry stopped hand gcode around 2000 (before 3d printing), but the language isn’t difficult and so you can do it.
Hand code is still used. Most frequently, a sub is hand coded with incremental moves and then called at a point. A lot of probing, weird moves, edge cases all still get hand coded.
Hand editing CAM produced programs constantly occurs at the machine as well.
3D printing gcode is pretty ugly because hobbyists “improved” the language without really understanding why it was so stripped down.
Modern Gcode is really meant to be machine generated and human readable/editable.
Couldn’t I at least punch it out on a card?
You could do that too. I’m not old enough to punch cards but that doesn’t need to stop you.