For the last half year to a year, the calorimeter group
has been living with two different result banks. The idea
behind supporting multiple banks was that we felt that we
didn't have enough information to know for sure what was
required for reconstruction, and we needed more experience.
The problem has been that both banks have not been supported
equally. Adherents to a given bank have worked on supporting
that bank, neglecting the other. Currently we have the
situation that only one bank or another can be filled, even
though the information is available for both. This has caused
many problems for internal and external collaborators trying
to work on analysis. An example of such a problem was when
our Korean collaborators were experimenting with electron
identification some time ago. They wanted to have the total
energy and inner and outer energy from the calorimeter for a
shower, plus have the moments of the energy loss distribution.
While this sounds reasonable, this was not (and is still not)
available, because the total energy of a shower is only
available in the ECRB, and the moments of the distribution are
only filled in ECHB (although they have been defined in both
banks). This is one of a number of problems caused by the
present situation.
I suggest that we now know enough about the analysis to define
a new bank, with a new name, and put the quantities of
mutual interest in it. Almost all of the variables in each
bank have been used in one analysis or another, and have
been found to have value. Some philosophical differences between
the two existing banks:
1) at what point do you match inner and outer calorimeter
information (before or after DC analysis)? Stepan has found
that you get more flexibility for photon reconstruction if
you wait until after the DC HBT analysis. This may also be
true for neutron reconstruction. On the other hand, for the
typical event containing one high-energy electron, it doesn't
make any difference, and it decreases the electron efficiency
(as in the next point).
2) do you define a hit using inner and outer separately,
or combine the information? Volker initially suggested, and
Stepan later confirmed, that you have higher electron
efficiency if you count the inner and outer together when
looking for the position information. (E.g., strip W34 is
hit if WI34 OR WO34 is hit)
I think what we need is to come up with a method where we
gracefully manage to keep both hit-level information and
the combined information. After tracking is performed and
electrons distinguished from pions, we could coalesce the
two types of information and eliminate a number of the
variables (like the moments of the distributions). This
second step probably already fits into Stepan's scheme.
I suggest we have a meeting to decide the final calorimeter
result bank format (at least the format for the next year
or so). I propose we meet next Monday at 11:00 am; a backup
time is Tuesday at 1:30 pm. I will find a room.
If you feel there are more issues than the two I mentioned
above, send them to the egn_calcom list so our discussion
can be more focused next week.
- Will