Attendance:  Kent Blackburn, Masha Barnes, Phil Ehrens,  Duncan Brown,  Teviet
Creighton,  Jolien Creighton,  Greg Mendell,  Lin-qing Wen,  Warren Anderson,
John Whelan,  and Joe Romano.


Discussin of Conventions document circulated to the group:
---------------------------------------------------------

Kent suggested that we should include DTT and DMT in Scope section.  (Get
Daniel and John on board.)   He also thought that references would be useful.
Might add something about the other alternatives.  Spell out equation (4).

The need for other metadata came up and it was emphasized that we have two
types of object.  Vectors versus series.  The latter includes metadata,  while
the former is mor internal to certain objects.  All of these things should
have some reference.   Numerical recipes seems to be the base reference.

Tev/Kent/John all prefer to pack things according to natural packing.  LAL
sequence versus time-series.  Kent suggested the notation that series and
things containing metadata are "exchange" level and should be consistent among
packages.  

Joe asked about definition of ensemble average.   Add statement that says
the method used to compute the ensemble average to the document.

People weighed in on one-sided versus two-sided PSD's.  Some wanted only one
type.  Others wanted both.  Currently the datacondAPI only produces two sided
objects.

Patrick proposed a telecon at 3pm CDT to continue this discussion.   Kent
agreed to have the Caltech operator initiate the call.   Patrick will phone
Sam to include in this afternoon's call


Change to LALWrapper:
--------------------

conditionData:  patrick explained how we got here.  If dataDistributor is
wrapper,  pass null pointer so that people wont get into a problem.   The
structure itself would need to be null.  Yeh,  that could be null in case
where datadistrbutor is wrapper.  

FCT code generates information that could have been communicated in
conditionData() function if there.   Right now,  this metadata is calculated
on every node.   

Make it null if load-balancing is enabled.   If someone specifies
datadist=search and load-balancing is true then exception.   But this is a
problem since applySearch can actually deal with dataDistribution and
load-balancing.  Kent is not opposed but can turn into blocking modes.   This
is difficult.   

Decision:  Will remove exception and will add NULL assignment to the code.
Duncan will modify LALWrapper based on Masha's change to include and
documentations should reflect that using this must not use rank or size when
doing load-balancing.  In any case, DLB will not happen soon. 


Stochastic DSO: No progress since last week.   Talked to warren and passed
baton onto him. 

Inspiral DSO:  no progress for this week.

FCT DSO:  no progress for this month.

Power:  no progress. 

Periodic:  
---------

Greg reported on this.  Wrote a basic thing to fold into routines.  Tev does
not think that amplitude modulation is in LAL.  Jolien ... not there yet.
David chin had a version from a long time ago which he is trying to bring up
to standard.   Prototypes have been decided.  Ask David Chin.  Greg will give
us a report next week.  


Other business: 
--------------

Phillip has nothing to talk about.  LAL package Comm, the documentation is out
of date.  Only about two paragraphs of text.  Sam raised problem report.

Agenda "work plan" for LAL and LALwrapper.   Need more people to start
attending the meetings. 

Next meeting:  Thursday,  21 June 2001 @ 10am PDT