Addition on computations in the interior.

git-svn-id: https://svn.dealii.org/trunk@4981 0785d39b-7218-0410-832d-ea1e28bc413d

diff --git a/deal.II/doc/tutorial/chapter-2.step-by-step/step-10.data/results.html b/deal.II/doc/tutorial/chapter-2.step-by-step/step-10.data/results.html
index 1693e85ea7e1fc72bb374ad94f26e67bcf964fcb..6ad2dc203a12d282dff7a3afcd3988caf0f3522e 100644 (file)
--- a/deal.II/doc/tutorial/chapter-2.step-by-step/step-10.data/results.html
+++ b/deal.II/doc/tutorial/chapter-2.step-by-step/step-10.data/results.html
@@ -74,7 +74,14 @@ with dashed lines the exact circle:
 </table>
 Obviously the quadratic mapping approximates the boundary quite well,
 while for the cubic mapping the difference between approximated domain
-and true one is hardly visible already for the coarse grid.
+and true one is hardly visible already for the coarse grid. You can
+also see that the mapping only changes something at the outer
+boundaries of the triangulation. In the interior, all lines are still
+represented by linear functions, resulting in additional computations
+only on cells at the boundary. Higher order mappings are therefore
+usually not noticably slower than lower order ones, because the
+additional computations are only performed on a small subset of all
+cells.
 </p>
 
 <p>