[Clas12_software] Software meeting minutes, July 24, 2014

[Mac Mestayer]

Hello;

Here are the minutes from today's meeting:

COMMON TOOLS PROGRESS: Gagik

- i/o done
- finishing up on new bank definitions
- Geometry Service; he and Jeremiah are working on a common
geometry api for all detectors starting with DC, EC, PCAL, FTOF.
Not only will it provide array lists of, for example, wire
mid-plane positions, etc., but it will contain many common
geometrical methods, e.g. line-plane intersection, etc.
The goal is to have the same "look-feel" for all sub-detector
geomety.
- PCAL/EC reconstruction is progressing.  The two detectors
are being treated together.  A prototype cluster-finder has
been written.
- Event Building: first step is DC track/ FTOF hit linking to
return particle beta and time of flight to various DC measurement
planes.

TRACKING - Veronique
- using ced to debug track-finding, fitting and also to investigate
track peculiarities such as large energy loss
- she measured track efficiency as a function of track angle for
5 GeV/c electron tracks in the mid-plane.  She sees efficiencies
of 95 - 98%; investigating cause of inefficiency.
- the measured 'cross' spatial resolution is now consistent with
the simulated wire-hit resolution after a bug was fixed.

BIT-SHIFT SEGMENT FINDER: Dave H.
- Dave has extended his (very fast) bit-shifting segment finder
to the job of linking segments from different superlayers.
He simply represents each segment as a bit in a new map
of six layers, each layer now representing the found segments
on one superlayer.  Because of the stereo-angle shifting of
track segments from tracks not at the mid-plane; he will create
separate maps for superlayers 1,3,5 and superlayers 2,4,6.
The idea is to tune the parameters for very high efficiency and
then to use it to pre-filter data before we implement TRAC's
noise pruners which are much slower.

HOW TO MODEL DETECTOR IMPERFECTIONS: Mac
- I reported that Johann and I are working on a data-base project
which would take time-stamped status tables for equipment (for
example, HV boards) and map them into a time-stamped status for
all wires.
- I then argued that we should not bother putting these dead areas
into gemc because it was sufficient to let the track reconstruction
throw out these wires whether or not they had a hit on them.
Gagik made the very good suggestion that this hit pruning job
should be done by a simulation post-processor, "gpp".

GOALS OF FALL REVIEW: Veronique, Volker
- We plan to have an internal software review mid to late September.
We talked briefly about its agenda.

 				that's all,  Mac

"mestayer at jlab.org", (757)-269-7252