From: Wolfgang Bangerth Date: Fri, 22 Oct 2021 23:11:39 +0000 (-0600) Subject: Update documentation of class Quadrature. X-Git-Tag: v9.4.0-rc1~903^2 X-Git-Url: https://gitweb.dealii.org/cgi-bin/gitweb.cgi?a=commitdiff_plain;h=refs%2Fpull%2F12866%2Fhead;p=dealii.git Update documentation of class Quadrature. --- diff --git a/include/deal.II/base/quadrature.h b/include/deal.II/base/quadrature.h index 22261e79e6..475229341d 100644 --- a/include/deal.II/base/quadrature.h +++ b/include/deal.II/base/quadrature.h @@ -33,16 +33,20 @@ DEAL_II_NAMESPACE_OPEN /** * Base class for quadrature formulae in arbitrary dimensions. This class - * stores quadrature points and weights on the unit line [0,1], unit square - * [0,1]x[0,1], etc. + * stores quadrature points and weights in the coordinate system of + * a reference cell (see the ReferenceCell class) and as such serves to + * represent quadrature points and weights on the unit line segment + * $[0,1]$ in 1d, on the unit square or unit triangle in 2d, as well as + * the unit tetrahedron, cube, pyramid, and wedge reference cells in 3d. * * There are a number of derived classes, denoting concrete integration * formulae. Their names are prefixed by Q. Refer to the list of * derived classes for more details. * - * The schemes for higher dimensions are typically tensor products of the one- - * dimensional formulae, but refer to the section on implementation detail - * below. + * At least for quadrilaterals and hexahedra (or, more precisely, since we work + * on reference cells: for the unit square and the unit cube), quadrature + * formulas are typically tensor products of one-dimensional formulas (see also + * the section on implementation detail below). * * In order to allow for dimension independent programming, a quadrature * formula of dimension zero exists. Since an integral over zero dimensions is @@ -53,17 +57,20 @@ DEAL_II_NAMESPACE_OPEN * of these formulae is their use in QProjector, which will create a useful * formula of dimension one out of them. * + * *

Mathematical background

* * For each quadrature formula we denote by m, the maximal degree of - * polynomials integrated exactly. This number is given in the documentation - * of each formula. The order of the integration error is m+1, that - * is, the error is the size of the cell to the m+1 by the Bramble- - * Hilbert Lemma. The number m is to be found in the documentation of - * each concrete formula. For the optimal formulae QGauss we have $m = 2N-1$, - * where N is the constructor parameter to QGauss. The tensor product formulae - * are exact on tensor product polynomials of degree m in each space - * direction, but they are still only of m+1st order. + * polynomials integrated exactly on the reference cell the quadrature + * formula corresponds to. This number is given in the documentation + * of each formula. The *order* of the integration error is m+1, that + * is, the error is the size of the cell to the m+1 by the + * Bramble-Hilbert Lemma. The number m is to be found in the + * documentation of each concrete formula. For the optimal formulae QGauss we + * have $m = 2N-1$, where $N$ is the constructor parameter to QGauss. The tensor + * product formulae are exact on tensor product polynomials of degree m + * in each space direction, but they are still only of (m+1)st order. + * * *

Implementation details

* @@ -73,11 +80,7 @@ DEAL_II_NAMESPACE_OPEN * one dimension. There is a special constructor to generate a quadrature * formula from two others. For example, the QGauss@ formulae include * Ndim quadrature points in dim dimensions, where - * N is the constructor parameter of QGauss. - * - * @note Instantiations for this template are provided for dimensions 0, 1, 2, - * and 3 (see the section on - * @ref Instantiations). + * $N$ is the constructor parameter of QGauss. */ template class Quadrature : public Subscriptor @@ -369,7 +372,9 @@ class QIterated : public Quadrature public: /** * Constructor. Iterate the given quadrature formula n_copies times - * in each direction. + * in each direction. The result is a tensor product quadrature formula + * defined on the unit hypercube (i.e., the line segment, unit square, or + * unit cube in 1d, 2d, and 3d respectively). */ QIterated(const Quadrature<1> &base_quadrature, const unsigned int n_copies); @@ -378,7 +383,9 @@ public: * defined by adjacent points in @p intervals in each direction. The resulting * quadrature rule will have `base_quadrature.size() * (intervals.size() - 1)` * quadrature points if no quadrature point of `base_quadrature` is positioned - * on the boundaries. + * on the boundaries. The result is a tensor product quadrature formula + * defined on the unit hypercube (i.e., the line segment, unit square, or + * unit cube in 1d, 2d, and 3d respectively). * * @note We require that `intervals.front() == 0` and `interval.back() == 1`. */