Remember that you never have perfect control over the exact shape of anything. You are constrained to what the material can achieve. If you could just draw out the chine, keel and stems the way you wanted them and have it work, you would not need to worry about whether the surface will develop.

I probably did not express what I meant very well. I agree that you don't have perfect control in either case, but if you define the stem through an intersection of the developable surface with a plane, you can no longer directly manipulate the shape of the stem. You can only change the shape indirectly by manipulating the chine extension. For a multi-chine hull, I am struggling to see how you would do that and keep more than c0-continuity in the stem. Of course, defining the curve of the stem first means that you give up control over the shape of the surface, and it may not be developable. Maybe there are CAD packages that have clever ways to do this?

My CAD package will display rulings for developable surfaces, but while I will look at them, I also look at sectional shapes and surface curvature and make an educated guess as to whether the plywood will be able to form to something resembling what I want. The analysis tools are useful, but experience is the final arbiter.

I have been looking at the sectional shapes as well, and the developable rulings do not look very promising to me, but I have no experience. The sectional shapes for non-developable rulings look much better to me, but then there is the issue of twist along the ruling lines. I wrote a parametric kayak generator program to analyze the developability of hulls with varying deadrise, flare, rocker, rake angles, fullness etc. to try to get some insight into what might be buildable and made the same observation over a range of reasonable (and unreasonable) looking designs: the mathematically perfectly developable rulings lead to very high curvature in cross sections because of slant in the ruling lines. The image below shows cross sectional shapes for a hull that is somewhat similar to the Shearwater Sport for rulings that have zero twist and three degrees twist. Compared to straight rulings, the volume in the bottom panel increases by 17% for the no twist ruling and by 6% for the three degree twist ruling. The curvature in the cross sections for the three degree ruling still seems high to me, but I have no idea how much curvature and how much twist the plywood can handle.

You can get very thin plywood which is probably the closest analog for scale models because the grain orientation does effect how the material bends.

I found some at aircraft spruce, not exactly cheap, though.