Attendance:  Kent Blackburn, Masha Barnes, Phil Ehrens, 
Alan Weinstein, Joe Romano,  John Whelan, Duncan Brown, Jolien Creighton,
Patrick Brady

Issues arising from the IUL MDC:
-------------------------------

The LDAS team is addressing issues that arose.  There were some problems that
arose with the mpiAPI.  From systems perspective did not have free hardware.
Something was restarted and it left the state of the node useg in the mpiAPI
confused. Working on something more robust.   Came across a problem of 4 jobs
reported on two nodes where there were no jobs running.  A matter of garbage
collection

If when LDAS is rebooted,  mpiAPI does a lam wipe at startup.  

How can we find out the logic of the mpiAPI?   Kent suggested that it may be
more meaningful to develop tests that check the requirements.  Phil would like
to have a summary of the mpi tcl file 

Next meeting we will deal with the testing and design of the mpiAPI.  Kent
thinks this meeting would provide an opportunity to design review.  

Duncan is going to address Sathya's bugs and try to verify his code using the
inspiral DSO.   Bruce has provided him with a set of chirps to inject and test
the coverage of the template bank.


Conventions:
-----------

There is a need for firm conventions in LDAS and related analysis software.  A
conventions document needs to be prepared and presented to the software
committee.   Issues to be addressed:

   Sign convention of the FFT in the datacond API.  Joe knows the sign convention
   is opposite.  Kent wants an absolute:

   datacond has -i  for forward FFT                                        |   how many actions rely on the FFT?  
   LAL      has +i  for forward FFT  /

Also add discussion of 1/N normalization in FFT's.  Circulate suggested
spectrum,  etc to everybody!

Stochastic DSO:
--------------

Joe/John are working on their code.  They have a local CVS repository they
have a stochastic package.   They copied power and are editing those
functions.   They have been through most of the headers and the .c files.
Have done a little bit on ConditionData.c   Will have something to put into
the CVS archive.   Can tar it up and put it into the LALwrapper CVS
repository.   

Alan will just start with getting the MPI MDC stuff fixed ........


Action Items:
------------
   1.  Testing and design of the mpiAPI.