@DanielPiker, Yes, that’s pretty close to what I want. I reduced my strength’s and then got closer to your results.
However, I would like to have a “true maximium” (rigid maximum) angle between the segments. For example the following curve (tangent match to the straight lines at the ends) which I divided into 3 x 90 (270) segments with ClampAngle 1 degree. Then I’d expect at least 90 segments (1/3 of total segments) at each end of the curve to have max 1º angle. But it seems the Solver converges at much less segements than that (~40 seg’s in the pciture), meaning that it converged at +2º degree (although max was set to 1º).
It would have been interesting to be able to have a rigid minimum 279º between two segments (almost straight) and 180º (straight) where possible, that is, where minimum angle isn’t enforced by necessity. Or in other words, “default = straight aligned segment pairs, and bend only the minimum number of segments at each end to its limit angle”. (minimum and maximum notion swapped here, the current notion is a bit confusing…).
The real world problem which I’m exploring is trying to autofit some rubber hoses/pipes/wires while keeping tangency at the curve ends. For this I would need to stay within a limited bend radius as to not damage the pipe/hose/wire.