Winding order for Open GL

[dean goodman]

I have a volume of data where I am making isosurfaces through the volume with a particular threshold. In plotting the vertex triangles which i discover from the possible 12 edges of a cube and rendering them in Open GL, some of the triangles remain dark with the shade model. I assume that it has to do with the winding order of the vertex. Computing the dot product and adjusting for negative normals does not solve the problem. Does anyone know a good resource to discover how to get the shade model correct with Open GL in PB or what the problem might be...thanks...