[[Imported from SVN: r2304]]

[[Imported from SVN: r2303]]

[[Imported from SVN: r2281]]

[[Imported from SVN: r2203]]

[[Imported from SVN: r2202]]

[[Imported from SVN: r2201]]

[[Imported from SVN: r2200]]

[[Imported from SVN: r2199]]

[[Imported from SVN: r2198]]

[[Imported from SVN: r2197]]

[[Imported from SVN: r2062]]

[[Imported from SVN: r2061]]

[[Imported from SVN: r2047]]

[[Imported from SVN: r2046]]

[[Imported from SVN: r2045]]

[[Imported from SVN: r2041]]

[[Imported from SVN: r2040]]

[[Imported from SVN: r2039]]

[[Imported from SVN: r2038]]

[[Imported from SVN: r2036]]

[[Imported from SVN: r2027]]

[[Imported from SVN: r2023]]

[[Imported from SVN: r2021]]

[[Imported from SVN: r2020]]

[[Imported from SVN: r2019]]

- do not hardwire the 3d material parameters
- add method to create a straight initial rod
- remove Neumann damping
- remove anisotropic damping
- do not hardwire the rod material and shape parameters
- use the C++ IPOpt, if requested
- use absolute corrections size for termination
- better determination of the overall convergence rate

[[Imported from SVN: r2018]]

[[Imported from SVN: r2017]]

make sure the fd hessian is symmetric; hardwire Dexp(0) for more speed; use symmetry when computing the differential of the geodesic interpolation

[[Imported from SVN: r2016]]

[[Imported from SVN: r2015]]

[[Imported from SVN: r2014]]

[[Imported from SVN: r2013]]

[[Imported from SVN: r2012]]

[[Imported from SVN: r2011]]

[[Imported from SVN: r2010]]

[[Imported from SVN: r2008]]

[[Imported from SVN: r2007]]

[[Imported from SVN: r1891]]

add a new method which averages out a dg function in order to obtain a p1 function.  This is used to pass Neumann data coming from a rod to the hierarchic error estimator

[[Imported from SVN: r1866]]

[[Imported from SVN: r1865]]

[[Imported from SVN: r1793]]