Peter,  Phil,  Kent,  Masha,  John,  Joe,  Warren,  Patrick,  Jolien, Tev,  
Linqing ......

How to drive LDAS:   

Layer I:  Peter has given some thought about driving the system.  What he sees
is a tcl library that provides a namespace ldasJob and functions inside that
namespace.   It would check environment variable for username; the would be
information about the job sent back in an array.  Then if you wanted to submit
two jobs in a row  there would be a couple of lines to do this .............. 

Then there will also be a utility which calls the library.  Job script which
will contain the list of parameters which are needed to execute the jobs.  One
magical thing that needs to be done is handle input files.  They can be given
to LDAS as a URL.  This is an alternative to pushing it onto LDAS for file
access.   Filename ...........

Layer II:  a queing system which allows submission of jobs and checking of
information for executing sequential jobs.  Will there be a logging layer?
Yes.   

Will this mechanism be able to understand whether the user 

Another thing that will come up eventually ........ 

Using external grid stuff:   needs additional work but would be able to fit
together into other things ......... both complement.   


11:30am Pacific on Monday for telecon between Peter, Scott,  Kent,  etc......

Looking at leaving both technologies in place for both ........ 

Would never expect the request manager to do whatever the user wants.  There
will always be a place for client side request information.   Grid request
manager will be there for data transmission and maintenance.   


Howto document changes have been made:  just waiting for libtool going into
ldcg so that the commands will work.   The web pages are up to date and ready
to go.

datacondAPI:  bug and memory leak.  This has been fixed for this release.  We
need the new release.   We'll get a new LDAS.  

Table design:   do we need these things frozen before the upper limit run?
Only in regard to testing the shared objects.   DSO developers need to repond
quickly.   Should we set a target date?   Yes.   1 December target date.   



Is there a table for storing spectra,  etc.   It has metadata and a mime-type.
There is also one 

Jolien is going to work on the document and table designs.   What other
components populate these table?   DMT,  datacond,  DSO's.   

What is the status of LDAS release?   If they dont have any more big problems,
it might be out the door tomorrow.   Target that next week we will discuss
changes and implications of new conventions.   


Stochastic-Bar:

they are trying to track down to the 


FCT:   


Masha has been having trouble running jobs on the test system.  They are
trying to find out what's going on.  Seem to be having trouble with MPIInit().
Duncan thinks he has seen something similar.   There was a wrapperAPI job fail
last night.   

Working towards getting ready to test existing FCT code for scientific
testing.   They are using Teviet's injection code.  They are trying to
understand how an FCT snr and the effective distance work out.   They are also
trying to understand if there is an easy and natural way to increase the
parameter space sampling.   First step to do what the matched filter does,
and then try to do better.   The natural FCT spacing is very wide in some
dimensions.   They are trying oversampling.   The main thinig that is the
hierarchical.   



John has a problem with building.