Versions Compared

Key

  • This line was added.
  • This line was removed.
  • Formatting was changed.

Status bits

Storage of data producer names instead of status bits

Starting at CalClusters level, objects are less numerous in the event (as
compared to lower level objects) and can be a little bigger. This is why,
for each data object, we would like to store the name of the producer as a
std::string, and remove the equivalent status bits. Such a mechanism
can adapt to any change in the algorithms, and be used with new
user algorithms, a lot more easily than status bits.

Avoid algo-oriented status bits

We should avoid status bits saying things about algorithms, because
this reduce our freedom with the way we arrange and change those
algorithms. We should prefer eventually status bits which say
things about the "state" of the data.

Status bits methods

For the useful status bits (for example the ones saying if this or
that data is valid), the methods signatures could be made more
uniform in the various CalRecon classes.

...

The current implementation is more or less hard coding that there are
two passes, and select the best energy candidate depending on the pass.
We propose something more flexible : there is no information about the
pass, and the best energy candidate is selected depending on the ones
which are present in the list : energy of CalValCorrTools if it is present,
and CalRawEnergyTool otherwise.

A detail: why duplicate the data fo the best CalCorToolResult ? Why not
simply point to it ?