PabloMack,
Each individual headstock has it's particular amounts of radial and axial runout. By making a toolholder on the headstock on which it will be used, the runout of that headstock can be significantly reduced. At least, that's what I think. (The TIR of my Beall ER-32 collet chuck differs significantly on my two headstocks 0.0006 inch on my mill and 0.0016 inch on my lathe.)
In mid-January 2020 when I was considering which Sherline lathe to buy, I asked Sherline if/when they would sell ball-screw lathes. The reply I received was, "Ball screw lathes will be available for sale later this month." It took a lot longer than half-a-month for Sherline to start selling them.
Knowing myself, if/when I go full CNC, it would probably be with machines with ball screws. Varying and changing amounts of runout on the different axis doesn't bother me when operating manually, but I have thought that I would find it intolerable with CNC to need to reprofile or readjust the amounts of backlash as they changed with usage of the machine.
I added the rotary column attachment to my mill, to allow the column to be turned to an angle for drilling at an angle, and to make aligning the column to the table easier (than using shims). I don't like this accessory, as it is difficult to realign the column. It doesn't turn smoothly, and the act of tightening its screws causes it to shift. I will probably remove the accessory, and keep the column always vertical. Keeping the column vertical would allow me to add a buttress to stiffen the column. Stiffening the column would make drilling at an angle via CNC-simulated-quill more doable. I agree, though, that mounting the workpiece at the angle is the better way
Karl