Task #3223: Cleaned up Demixer.cc, removed boost::multi_array dependency, several other source files cleaned up and documented.

Task #2942: We need to set 'outputs' field 'ok' when we're exiting early because our output file already exists.

Task #2942: Do not let the DPPP recipe clobber its output by default. This is needed if we want to do a re-run to process missing Measurement Sets.

Task #2942: The target pre-processing pipeline now also creates different instrument- and sky-models for the DPPP and BBS run.

Task #2942: Reverted commit r21701. Don't know if it's really needed, and it doesn't seem to work the way I thought

Task #2942: Updated dppp node script; now uses temporary MS-file as working copy (if needed); skips processing if output file already exists.

Task #2942: Create separate sourcedb and parmdb for DPPP and BBS run, instead of overwriting them. This is needed when we want to re-run part of the calibrator pipeline.

Task #2942: We also need to store self.outputs in the statefile. Otherwise we won't get the correct outputs back for a skipped recipe.

Task #3446: Keep memory usage under control for beam formed data, and explictly list what flat memory is used for

The SQL function must be added to the database and the OLD SETSCHEDULE FUNCTION MUST BE REMOVE MANUALLY!
The C++ interface changed from ptimes to strings.
The Java interface remains unchanged.

Task #3250: Include the new ApplCommon/LofarDirs.h header file when using any of the LOFAR_*_LOCATION preprocessor variables.