Then it would appear that a completely different mindset is needed for 3D panels, from that needed for 2D panels.
With a 3D panel, whether it has tucks, shelves or broadseam, one designs one's panel with a certain amount of camber, then builds it and gets the amount of camber one has designed, within close enough limits for a junk sail. Moreover, that camber is present, not only at the mid-height of the panel, but also, to a reasonably even degree, over more of the height of the panel. Of course, the mid height of the panel is blown out into something of a curve; but overall, one gets the camber one is hoping for, which is something close to a trapezium in cross section. There is no question of "pulling harder" or "less hard", to adjust the camber, because this will have no effect. Light tension, to remove the slack, is enough, and more tension will have no further effect.
I would be alarmed if my sail started pulling into a concave shape at luff and leech, putting more depth into the middle of the sail. This is something over which I would have no control, and would rely totally on the stretch characteristics of the cloth. It is the kind of thing that happens if one pulls a handkerchief out by its four corners, and then drops an orange into it. One gets some camber, but the amount and position of that camber seem to be in the lap of the gods.
Again, I need to emphasise that I have nothing against the 2D, barrel cut panel. It has its place in the scheme of things, which is to help first time sailmakers get onto the bottom rung, and to enable others to play the 80/20 rule, getting the most return for the least investment of their time and skill. For others still, well, the 3D panel seems to me to be worth that extra bit of investment.