OFF-LINE DC HISTOGRAMS:
Richard presented off-line analysis histograms -
containing raw and Hit Based tracking information.
SOFTWARE STANDARDS:
A long debate occurred over where to
place general purpose monitoring histograms for RECSIS. A
consensus emerged that histogram filling and booking
routines should not be in any of the major RECSIS
subsystems DC, EC, TRK etc... or in the default
RECSIS - and that the histogram filling and booking
routines already in RECSIS should be removed. However,
people also felt the need to put standard monitoring
and debugging histograms under cvs control and
to provide mechanisms for passing information to
histogramming routines that might not be available from
BOS banks.
To achieve the goal of both removing histograms
already in RECSIS and to provide a structure in cvs for
storing and developing histogramming packages a three-fold
strategy was proposed.
1) USER routines can attach to common blocks in
other packages for the sole purpose of generating
histograms for monitoring/debugging purposes
2) One or more cvs directories will be established
which will contain user plug-ins for RECSIS -
the exact organization of which is up to
individual developers - a model of this already
exists in packages/ntuplemakers, presumably
a new library will be set up for off-line
monitoring histograms
3) As a last resort, one can insert a dummy new USER
subroutine into a package which in the
specialized version of RECSIS monitors a
variable that is lost or overwritten before
the regular dummy USER routines are called.
To restate the essence of the above arguments,
various copies of the user routines will exist in the
cvs repository, by default, the dummy user routines
will be used when one compiles RECSIS (just like always).
However, one will be able to also compile RECSIS
with the alternate (smart) user routines, creating
monitoring or analysis versions of RECSIS. Some the
RECSIS rules have been loosened for the USER routines
to make it easier to create histograms within the package.
The point of this exercise to remove histograms
from the basic RECSIS package by making it easy to
put histogram calls in the individual's user area. Let
me note that the various libraries STILL must
communicate between packages by BOS banks. The philosophy
here is that the USER package is an exception since it
exists solely for software development and histogram
generation - Note again that in the official version of
RECSIS the USER area contains only dummy routines
(no histograms no anything).
Richard and Francois will implement these changes
to RECSIS and remove histograms already in RECSIS packages.
RECSIS DOCUMENTATION:
RECSIS documentation was discussed, some
improvements were suggested, including consolidating
the documentation on TCL variables - However there was
no volunteer to update the web documentation(14 steps to
Nirvana) - for now we will have to live with it. Arne
may have some supplemental RECSIS documentation that is
more up-to-date.
OFF-LINE SHIFT STATUS:
Will Brooks presented his vision of the Off-Line
shifts. Various discussion points included:
- RECSIS on the Sun's runs 2-5 times slower than
the other platforms - the bug is known, but needs
to be fixed if we wish to accumulate statistics for
analysis.
- Resources available for off-line analysis,
including the batch farm, and a large amount of
disk space
- People are needed to run off-line shifts
- Various schemes for getting to the data as
quickly as possible
- Mac felt that a primary responsibility of the
offline group was to verify the integrity of the
information in the BOS banks by reading the ROC banks
directly - however stand alone software only exists
to decode the ROC banks for the drift chambers -
producing a stand alone program to regenerate all
of the banks (EC, SC etc..) from the ROC banks would
be a major effort. It is unclear what the rest of the
community feels about such an effort.
NEW DC ROADS FILES:
Francois is in the process of generating new roads
files currently he is planning on generating road for the
following run configurations.
Torus Mini-T - fraction of Maximum current
.25 .75
.50 .75
.75 .75
.90 .75
-.25 1.00
All runs will have roads for positive and negative
tracks over a momentum range of .1-5.0 GeV and with full
angular acceptance (phi -30->30 deg, theta 8-130 deg) and
using the new field map file (correct toroid geometry).
Francois is leaving for France next week, if any additional
files need to be created to reflect anticipated running
conditions please let him know as soon as possible.
NEXT MEETING:
There will be a special off-line meeting next
week Tue at 1:30PM at a place to be announced - the
purpose of which is to discuss preparations for the
upcoming run.
-Joe