Marek: Is the efficiency for all tracks or for stiff tracks? It is for
all primary tracks.
Karel: Is the fake rate 100% when you work within one sector? Yes. How
many would disappear if you do the merging with the other sectors?
Some, but the not very much.
Marian: Do you use the ADC amplitudes? Yes. Could a log. transform
help? Probably, it has to be investigated.
Lachezar: Do you have ExB effects in the simulation.? Yes, on the wires, not in
the volume.
Marek: Do you want to use HLT for low PT tracks? For the moment it is
used only for compression.
Dieter: We are showing the concept and what we can do now.
The timing for reconstructing a pile-up event is so small that we can
reconstruct on-line all tracks, secondary and primary ones.
The basic idea is to have the track seeds and fit then clusters. The
Hough transform is not crucial, it is one of the methods we can use.
Marian: What do you plan to use for secondary tracks? The Hough
transform is one good candidate.
Karel: Do you classify the cluster with respect to their width? Not
for the moment. It should be done with real data. In any case it can
help when searching for overlapping clusters.
Marian: How sensitive is the algorithm to the fake tracks? Since we are
looking for local max., it is sensitive.
Federico: Is the fake tracks the limiting factor? Yes, this is the most
urgent problem.
Basanta: What is the physics goal of this exercise? It is to show
that we have some algorithms ready for the real triggering. One needs
the track reconstruction in TPC for many physics triggering tasks.
Dieter: The e+e- pre-trigger gives you several handful candidates and
the we work with a part of data.
Marian: The efficiency. was only for primaries, wasn't it? Yes, it was, but
for the trigger the secondaries are not very important.
For primaries the momentum we can reach is 120MeV, but with standalone
tracking one can go below.