Summary of the meeting at BNL on September 23, 1998
Present: Flemming, Alv, J. H., Kris
I. Discussions on progress on Mock Data Challenge to date
1. summary of results - many segmentations faults in
reconstruction code.
2. discussion on how to get code to go through simulated events
3. What to do for the rest of MDC1.
a. Decide to treat the files generated as RDO's(Reconstructed Data
Objects).
To simulate generating PhDs (Physics DataSets), will simply read
the
RDO's, reject events
with no tracks and write the same format file as PhDs
b. Try to put together routine to combine tracks for MIDS. This
already exists in some form as copied from FORTRAN Sonata, but has
not been checked or debugged.
II. Discussion on Documentation
1. BRAT Manual (plan to have draft before coll meeting)
a. Organize manual so that it begins with an overview and goes
into progressively more detail.
i. How to use what is there without adding anything new.
ii. Tutorials
iii. How to add an analysis module
iv. How to add new data structures.
b. Flemming puts together appropriate documents that already
exist into the framework of Kris's
c. J. H. will work on writeup of later datastruct's
d. Alv would contribute to the discussion of analysis modules.
III. Discussion on Data Structures
1. DST
a. J. H. will propose a DST data structure by the collaboration
meeting.
It will probably be in the form of a rough BRAT .h file.
2. BRAHMS Tracks
a. Current Track structures were essentially taken from FORTRAN
Sonata and adapted to C++, then adapted to BRAT tables. In
the
meantime all
kinds of cross references made by pointers was added for
short term
simplicity, but makes the overall structure into a mess.
b. There is also a proliferation of track classes that have
essentially the
same functionality.
c. Discussed strategy for the Track classes in the framework of
BRAT
d. Decided to look into a track structure in which there is not
such a
proliferation of track classes.
e. Will come up with different track classes for FS and MIDS as
the tracks from
the two arms will not be the same.
IV. Corrections to BRAT
1. Clear clones arrays before exiting TPC and DC tracking
routines. This eliminates segmentation faults caused by the user
deleting classes outside
of these routines that are still held by the clones arrays.
2. Found that T1 is offset by 2cm in y which caused virtually no
tracks to be matched between T1 and T2. The tracks that were matched
were wrong
because of the offset. Implemented a workaround to use for the
remainder of
MDC1.
minutes by Kris edited by Flemming;
-- ------------------------------------------------------------------Flemming Videbaek Physics department, Brookhaven national laboratory Upton L.I., NY11973, USA
Phone: (516)-344-4106 FAX: (516)-344-1334 secr: (516)-344-7681 e-mail: videbaek@bnl.gov
This archive was generated by hypermail 2b29 : Tue Feb 01 2000 - 20:35:19 EST