From: bangerth Date: Tue, 22 Nov 2011 14:40:01 +0000 (+0000) Subject: Augment documentation. X-Git-Url: https://gitweb.dealii.org/cgi-bin/gitweb.cgi?a=commitdiff_plain;h=f81348bf6bd399d0f808f0c66de11eb88c4957c4;p=dealii-svn.git Augment documentation. git-svn-id: https://svn.dealii.org/trunk@24757 0785d39b-7218-0410-832d-ea1e28bc413d --- diff --git a/deal.II/include/deal.II/numerics/vectors.h b/deal.II/include/deal.II/numerics/vectors.h index dfd428656b..465dc991c0 100644 --- a/deal.II/include/deal.II/numerics/vectors.h +++ b/deal.II/include/deal.II/numerics/vectors.h @@ -310,8 +310,8 @@ class ConstraintMatrix; * that requires the computation of a gradient, then the provided * function is automatically projected along the curve, and the * difference is only computed on the tangential part of the - * gradient, since no information is available, on the finite - * dimensional one, on the normal component of the gradient. + * gradient, since no information is available on the normal + * component of the gradient anyway. * * * All functions use the finite element given to the DoFHandler object the last @@ -1720,7 +1720,20 @@ namespace VectorTools * weighting function is given, * i.e. weight=1 in the whole * domain for all vector - * components uniformly. + * components uniformly. Note that + * one often wants to compute the + * error in only one component of + * a solution vector (e.g. for the + * pressure in the Stokes system, + * when the solution vector also + * contains the velocity components). + * In these cases, the weight should + * be a mask, i.e., be a + * vector function for which individual + * components are either zero or one. + * This can easily be achieved using + * the ComponentSelectFunction + * class. * * It is assumed that the number * of components of the function