Changing template-depth back to 40.  Adding initialisation functions and constantProperties construction.

Renaming BinaryElasticCollisionModel to BinaryCollisionModel - some models that redistribute internal energy are not elastic.

Forgot to git add new files referred to in commit 5fca1e87 message.  Initial DSMC implementation now included properly.

Finished DSMC code initial layout and test solver - both compile.  Using lagrangian/dsmc local version of WallInteractionModel until requirements stabilise. Removed use of InjectionModel - not suitable, designed for continuum cases, requires single constProps from trackData - multispecies DSMC requires a List of constantProperties, one for each species.

Added distance to and index of closest surface patch to indexedVertex, using this to preform tests of near surface grading and refinement.

Fixes to point insertion, making faceAreaWeight normalised, not hard coded to cell size, adding tweak to fertex removal criteria to allow selective insertion at short-edge midpoint if the vertices have not already been flagged for removal.

    - still not consistent for pd as does not include buoyancy effects

while(runTime.loop())
{
...
}

idiom.

Demonstrated in the shallowWaterFoam solver.

Added point removal for short Delaunay edges in rotational controller as well as small and highly conntected dual cell removal.

- track previous time index and to avoid recalculating averages.
- make sure averages are up-to-date before calling write

- added end() method to functionObject, functionObjectList & associated classes
- moved outputFilters from src/sampling -> src/OpenFOAM/db/functionObjects

- this (now deprecated) idiom:
      for (runTime++; !runTime.end(); runTime++) { ... }
  has a few problems:
    * stop-on-next-write will be off-by-one (ie, doesn't work)
    * function objects are not executed on exit with runTime.end()
  Fixing these problems is not really possible.

- this idiom
      while (runTime.run())
      {
          runTime++;
          ...
      }
  works without the above problems.

Added point insertions on long edges for rotational controller and identified possible point removal candidates.

even when the dot-product of the Reynolds stress and the gradient of velocity is
negative.  Currently this is done by taking the magnitude, another alternative
would be to take only the positive part.