<div class="gmail_quote">On Sun, Aug 14, 2011 at 03:23, David Ketcheson <span dir="ltr"><<a href="mailto:david.ketcheson@kaust.edu.sa">david.ketcheson@kaust.edu.sa</a>></span> wrote:<br><blockquote class="gmail_quote" style="margin:0 0 0 .8ex;border-left:1px #ccc solid;padding-left:1ex;">
For plotting purposes, linear interpolation is fine. There are other things we could do with higher order interpolation, but they're not critical yet.</blockquote></div><br><div>I'm less concerned with the order of the basis functions than with the possible presence of a nonlinear reconstruction. For any order polynomials, interpolation is still a linear operator (could be written as a matrix).</div>
<div><br></div><div>If instead the interpolation involved a TVD or WENO reconstruction, or if it involved coarse level basis functions that were upwinded using a local characteristic decomposition, then it could not be written as a matrix and there would be extra semantics associated with transferring solution values versus increments or residuals.</div>