JLab Software Meeting
August 22, 2005
David Lawrence



Attending: David Lawrence, Elliott Wolin, Mark Ito, Steve Wood,
Rob Feuerbach, Dennis Weygand, Sergey Boiarnov, Vardan Gyurjyan,
Ed Brash (VRVS but cut off shortly into meeting)

Hall-A Analyzer
----------------
Rob Feuerbach gave an overview of the Hall A C++ Analyzer. The
slides were from:

 http://www.jlab.org/~feuerbac/halla/Podd/podd_summary.pdf
 
 The Analyzer is a ROOT based analysis package. It defines a
 set of base classes with virtual methods such as "Decode" and
 "Reconstruct" which are used to advance through stages of the
 analysis.
 
 Using ROOT's CINT, the analyzer can be easily scripted or even
 run interactively using "root". One can also extend the core
 analyzer by loading in shared libraries, even while in root.

Two configuration files are used to control the analyzer. One, the
"OdefFile" or output definition file is used to create histograms
and a TTree and fill them using global variables. The OdefFile
is parsed inside the Analyzer which translates variable names
into actual variables. In this way, the variables are treated
as read-only.

The second configuration file is the CutFile, which is used to
define cuts that either be used in histograms, or that can be used
to filter events completely out.

The Hall-A Analyzer reads in CODA formatted data. It can also be
used to write out selected events to a filter file. The event
buffers are written out exactly as read in so reconstruction must
occur again evey time the filtered file is read in. The
reconstruction rate is speedy enough though, that this is not
considered a limitation.

Calibration constants are currently stored in files, though they
have discussed moving to a database.


Online Alarms
---------------
A brief discussion was had on software alarm systems. It was
suggested that the alarm system used by CODA be made more
general so that it could be used for all systems. Specifically,
that CODA alarms should be displayed in the same window as
other systems, such as EPICS based ones. It was suggested that
the DAQ group would be the natural place to supply this type of
tool since it would be used in all halls.

There were some concerns though about how a well integrated
such an alarm browser would be with the EPICS controls.
Specifically, in Hall-B, when an EPICS alarm occurs, one can
click on the alarm window and get back to the control window
for the channel with a problem. It would take a lot of effort
to reproduce this functionality.

This topic will need to be discussed futher at a later meeting.


Next Meeting
--------------
The next meeting will take place in approximately 2 weeks on a
date TBD. Steve Wood will talk about the Hall-C analysis package
at the next meeting. Time permitting, David Lawrence will
talk about the Hall-D package currently under development.