Very nice! I played a bit with you CSG script to better understand what happens... Turns out that reducing resolution a lot (I used 16) allows to manipulate the draft much faster.
Then I compiled the result, and... let's say that triangulation is far from perfect. Here is the result after exportation in Meshmixer. The blue lines shows mesh issues (I guess T-junctions).
Apart from that triangulation issue, probably better result would be obtained if the CSG sphere had no distinct poles (something like the LDraw sphe primitive.
Then I compiled the result, and... let's say that triangulation is far from perfect. Here is the result after exportation in Meshmixer. The blue lines shows mesh issues (I guess T-junctions).
Apart from that triangulation issue, probably better result would be obtained if the CSG sphere had no distinct poles (something like the LDraw sphe primitive.