In the present version of EC, there are four levels of information
flow.
(a) Stripes: Individual ADC which is greater than the programed
threshold.
(b) Peaks : clustered stripes in each (layer, axis)--1 dimension
cluster ::ecfit_strips_peaks.F
(c) Hits : Combination of peaks, which satisfied the dalitz rule
--2 dimension cluster. ::ecfit_peaks_hits.F
This imply one needd to have at least one peak in each view
to have a hit.
(d) Shower : Combination of inner hit and outer hit. The
subroutine ::ecfit_hits_showers.F
This imply that one needs to have at least one inner hit and one
outer hit to have one shower.
For those cases that 1 inner hit and 2 outer hits or vise versa. The
ecfit_hits_showers uses some algorithm to pick the most likely
pair of as a shower. Throwing away hit information at this point
means that a different combination is ruled out. Of course, we can
always reanalyze the data. Then, it will be a case of storage cost
versus processing cost.
If the purpose of ECRB bank is to post reconstructed information for
other packages, the current ECRB bank is probably adequate for this
purpose. Saving intermediate result in ECRB
may just confuse the rest of world. It might not be a bad
idea to save "the information that some of us think may be handy to
have for the next level reconstruction" in a seperate bank.
The bank design is really highly coupled to the detail of the
reconstruction algorithm. Whether one should take the risk of wasting
storage capacity or process capacity probabely depends on personal
preference.
--
T.-Y Tung (:
College of William and Mary
Trailer City rm 14
Tel: (804)249-7541