Stephen R. Aylward wrote: >I think Luis' last point - that we pass itkVectors is nearly a >requirement, not just a preference - to keep the API consistent >looking. Right? > That's right, So far ITK users don't need to be aware of the fact that vnl is being used inside. The new itk::Array is a way of avoiding to expose a (vnl_vector) or an (std::vector) in the API. Luis