Buffer was dynamically allocated based on first entry of data type.
In case the second entry (e.g. Y) was larger, it was truncated.

Note: This needs to be variable in future!!

- The basic class 'data_obs' now contains variables to hold a
  runtime-defined number of data fields and supplementary channels.
- Many of the get/set methods for data and supplementary channels
  are now moved to the basic class data_obs
- Same is true for methods to calculate median/mean etc.
- Specialized data_classes may contain additional, specialized methods
  for setting/getting data,....

- Debugging info (commented) was added for Filter_Obs
- Faster start of PLASMON filter, as filter2 starts directly
  as soon as enough data are available from filter1.

String_Parser will produce string from generic class data_obs,
and call generic data receive methods only. The proper method
is selected by overwritten subclass method, e.g. data_obs_vector.

- Add generic methods to return data and supplemental data
- Channel needs to be specified in single calls
- Missing data value moved to class data_obs
- Introduce method to get data with char identifier (IN DEVELOPMENT)

Class String_Parser adds functionality to specify
a format string that is used to customize the
data records. For time, strftime is used, otherwise
fprintf-syntax is used.

Get the time from data-object represented as string
based on format specifier using strftime

A default value can be specified for Parameters_Reader.
This is useful if some parameters are not specified in
the file.

Set default values for calibration constants that
will be used when unspecified in parameter file.

yday in struct tm_time had to be set by
calling mktime

Previously defined entries of par-file can be
referred to with ${NAME}, and they will be
replaced accordingly when parsing the par-file.

Wait to start filtering until .5s of next second has
just passed.