The speed increase cause by the patch is, alas, hardly measurable.

[[Imported from SVN: r8360]]

[[Imported from SVN: r8353]]

[[Imported from SVN: r8352]]

[[Imported from SVN: r8351]]

[[Imported from SVN: r8350]]

[[Imported from SVN: r8349]]

[[Imported from SVN: r8348]]

This is now mandatory since Quaternions are not implicitly casted to Rotations
anymore (via Quaternion --> FieldVector<4> --> Rotation).  This implicit cast
lead to a nasty bug here, because it involved a normalization of the argument.

[[Imported from SVN: r8347]]

[[Imported from SVN: r8346]]

[[Imported from SVN: r8345]]

This is safer, because calling the constructor implicitly
can lead to unwanted quaternion normalization.
(happened to me...)

[[Imported from SVN: r8344]]

Even though multiplication of Rotations is the same thing
as multiplication of Quaternions, in the first case the
output is a Rotation class, and in the second on a Quaternion
class.

[[Imported from SVN: r8343]]

[[Imported from SVN: r8342]]

[[Imported from SVN: r8335]]

[[Imported from SVN: r8334]]

[[Imported from SVN: r8333]]

the fe functions on two different refinements of the same base grid.

[[Imported from SVN: r8331]]

[[Imported from SVN: r8330]]

[[Imported from SVN: r8329]]

- set mu_c to 0
- introduce the shear correction factor 'kappa'

[[Imported from SVN: r8328]]

[[Imported from SVN: r8327]]

[[Imported from SVN: r8326]]

[[Imported from SVN: r8310]]

Use the long form for the quadratic energy stiffness.  I tested it: when kappa=1 is does the same as the short form

[[Imported from SVN: r8308]]

[[Imported from SVN: r8307]]

[[Imported from SVN: r8306]]

[[Imported from SVN: r8305]]

[[Imported from SVN: r8304]]

[[Imported from SVN: r8303]]

to set the shear correction factor.

[[Imported from SVN: r8290]]

[[Imported from SVN: r8277]]

[[Imported from SVN: r8276]]

[[Imported from SVN: r8261]]

[[Imported from SVN: r8260]]

[[Imported from SVN: r8259]]

[[Imported from SVN: r8258]]

[[Imported from SVN: r8257]]

choose the correct quadrature order for the matrices that create the norms to measure the discretization error

[[Imported from SVN: r8243]]

[[Imported from SVN: r8242]]

[[Imported from SVN: r8241]]