https://tams.informatik.uni-hamburg.de/research/3d-printing/nonplanar_printing/index.php
Subtractive CAM has had smooth/finish processing “forever” — I hope this makes it to at least one of the main slic3r branches soon.
I’m curious about the impact of the width of the shoulder around the nozzle on surface finish. My first instinct was to turn it down to get it out of the way of the path, but then I realized that would make a smaller radius and might be a worse surface finish. 
Thanks for sharing! Maybe limited in application in plastics but rather useful in ceramics for geometric tiles. The nozzles in clay printing are also much longer and narrower increasing your useful height. I’ve been doing non planar printing with Grasshopper for ceramic and indeed the nozzle is an issue when traveling downwards at an angle. Keeping the nozzle rim thin improves the finish. Upwards is no problem.
With little modification to the slicers, we could probably use variable layer heights (must specify a minimum layer height) so that we can print the slope for that individual layer that way.
As a future adjustment to the modification, we could slightly decrease the previous layer in spots where the desired height is less than the minimum layer height for the current layer.
I know most printers have a cooling fan at or near nozzle height. I have never had one nor have I seen a real need if you have your parameters set correctly. This being said, I seem to have more clearance than most people and think that rethinking the nozzle arrangement could help with clearance.
One benefit of my alternative is that clearance does not matter.
That’s a really cool technique with some great looking surfaces.
Has anyone put the printhead on the end of a robot arm so it could tip over by as much as maybe 90 deg on outside (convex) surfaces?