> The reason of the question is that in the TensorFunction the > dimensionality of the function is restricted to the dimensionality of > the embedding space.
That's the definition we use for a tensor: it's dimensionality is given by the embedding space. Everything else is just a vector. So in your case, I see two options: 1/ You use a Function<3> with 2 components 2/ You use a TensorFunction<1,3> that returns vectors whose z-component is always zero. In fact, this is what I would probably do: magnetization is inherently a dim-dimensional vector (i.e. a tensor) and the fact that your model only allows a magnetization with a z-component of zero, then just setting this component to zero appears natural. Best W. ------------------------------------------------------------------------- Wolfgang Bangerth email: [email protected] www: http://www.math.tamu.edu/~bangerth/ _______________________________________________
