We are looking at geometric data libraries to integrate with our software and are interested in any comparisons you can make between OpenCascade and SMLib. I'm sure many of you went through a similar decision process and am wondering what made you go with OpenCascade and how happy you are with it.
In order to give you a perspective on what our needs are, here is a summary:
We are a manufacturer of numeric control machines including milling machines and coordinate measuring machines (CMMs), and we provide the software to support them. For the milling machines, we provide software for manipulating 3D surface data and generating cutter paths across it. Most of our data comes in the form of IGES files. We get B-spline surfaces, parametric spline (power series) surfaces, ruled surfaces, surfaces of revolution and tabulated cylinder surfaces. Frequently, they are trimmed with outer and sometimes inner boundaries. We are interested in converting these to NURBS surfaces and in tesselating into triangle meshes. I know SMLib has extensive NURBS capabilities and am wondering how OpenCascade compares in this regard.
For our CMMs, we will be dealing with DMIS data. We need to take sampled points and convert them to features such as cylinders, cones, spheres, etc., and need to compare against nominal data. The Canonical Recognition "Advanced Algorithms" add-on for OpenCascade sounds like it might be useful.