<p dir="ltr" style="margin-top:0; margin-bottom:0;">You could also try the hierarchic Lobatto basis. It is still work in progress, though. There, the linear basis functions are numbered first, then the others. (We can talk in more detail if you want to know more.) It supports also higher order dofs. Those are ordered currently by subentity, but i think about an ordering by polynomial degree as well. </p>
<br>
<p dir="ltr" style="margin-top:0; margin-bottom:0;">Best, Simon </p>
<br>
<p dir="ltr" style="margin-top:0; margin-bottom:0;">Von meinem Xperia Smartphone von Sony gesendet</p><br><br>---- Lasse Hinrichsen-Bischoff schrieb ----<br><br>Dear dune-functions community,<br><br>I'm trying to implement a preconditioned hierarchical error estimator.<br>My ansatz space is P1 and I want to estimate the error in Q=P2 such that <br>I have Q = S + V, where S = P1 and V are the quadratic bubble functions.<br><br>It seems the HierarchicalLagrangeBasis is what I would want to use.<br><br>My question is, how do I know which indices correspond to the quadratic <br>elements and which are piecewise linear? The only thing I came up so far <br>is looking into the code of the local FE and noticing that the local <br>indices 0, 2 and 5 correspond to the linear parts and 1, 3 and 4 to the <br>quadratic ones (in 2D). Is it okay to use this fact and construct index <br>mappings from here?<br><br>Best,<br>Lasse<br><br>_______________________________________________<br>dune-functions mailing list<br><a href="mailto:dune-functions@lists.dune-project.org">dune-functions@lists.dune-project.org</a><br><a href="https://lists.dune-project.org/mailman/listinfo/dune-functions">https://lists.dune-project.org/mailman/listinfo/dune-functions</a>