Previously, the actual projection happened by calling the constructor
of one of the classes that implements points on a manifold.  This patch
introduces an new method projectOnto, which makes the projection more
explicit.

This is needed for the new gradientflow application.  The standard 'distance' method
doesn't cut it, because it is not differentiable near zero.  Therefore, even the
differentiable expression 'distance*distance' will fail to be differentiable for
an automatic-differentiation system.  I think in the long run we should replace
'distance' by 'distanceSquared' everywhere it is used.

Unfortunately, this patch is hacky: it only adds the method for the double/adouble
combination.

Which is the type returned by the derivativeOfProjection method.

[[Imported from SVN: r10067]]

This is needed for projection-based finite elements.

The code in this patch has not been tested yet.

[[Imported from SVN: r9887]]

[[Imported from SVN: r9886]]

[[Imported from SVN: r9588]]

We normalize unit vectors again in the constructor and the assignment operator.
This makes sure we never drift away from the unit sphere, and it also allows
us to init unit spheres with any value in R^n and be sure we obtain a unit
vector.  This makes the test pass again.  Leaving the projection out didn't
really make a measurable difference anyway.

[[Imported from SVN: r9574]]

This effectively means that we use another prolongation of the distance 
function on M into the surrounding space.  Since the prolongation does not
matter this patch should not change the algorithm behavior.  However, it
shaves off a few norm calculations and division.  I cannot really measure
any difference though.

A possible effect of this is that while all values should remain on the
manifold, they may start to 'drift away' due to numerical artifacts.
So we may have to add an occasional renormalization step eventually.

[[Imported from SVN: r9558]]

[[Imported from SVN: r9449]]

[[Imported from SVN: r9448]]

I can't make up my mind whether I want 'ctype' or 'field_type'

[[Imported from SVN: r9447]]

When using the classes with ADOL-C, T becomes an 'adouble',
which is a non-trivial type.  However, the convexityField
is constexpr (and with a reason -- otherwise we could write
the initialization together with the declaration), and
constexpr and non-trivial types don't work together.

[[Imported from SVN: r9444]]

[[Imported from SVN: r9397]]

And use it in the test.

This patch uses the new 'constexpr' macro.  Hence it is C++11 only.

[[Imported from SVN: r9195]]

[[Imported from SVN: r8369]]

It makes it difficult to be sure that an explicitly
requested type is really handed over to every place.

While we are at it: reverse the order of the template
parameters.  Now first comes the type, then the dimension.
This is more like FieldVector and friends, and hence
should be less confusing.

[[Imported from SVN: r8148]]

[[Imported from SVN: r8145]]

[[Imported from SVN: r8034]]

[[Imported from SVN: r7374]]

[[Imported from SVN: r7313]]

[[Imported from SVN: r7225]]

[[Imported from SVN: r7205]]

[[Imported from SVN: r7196]]

[[Imported from SVN: r7195]]

[[Imported from SVN: r7134]]

[[Imported from SVN: r7090]]

bugfix in thirdDerivativeOfDistanceSquaredWRTFirst1AndSecond2Argument.  Now it produces the same result as the fd approximation at least in for S^1

[[Imported from SVN: r7066]]

[[Imported from SVN: r7053]]

[[Imported from SVN: r7052]]

[[Imported from SVN: r7051]]

[[Imported from SVN: r7050]]

[[Imported from SVN: r7049]]

[[Imported from SVN: r7048]]

[[Imported from SVN: r7033]]

[[Imported from SVN: r6501]]

[[Imported from SVN: r6496]]

[[Imported from SVN: r6450]]

[[Imported from SVN: r6437]]