Attendees: David Adams, Vito Di Benedetto, Giuseppe Cerati, Lynn Garren,
Krzysztof Genser, Herb Greenlee, Alex Himmel, Tom Junk, Kyle Knoepfel, Katherine Lato, Gianluca Petrillo, Paul Russo, Erica Snider, Tracy Usher, Brett Viren, Hans Wenzel, Tingjun Yang

Project Report
--------------
by: Erica Snider

Tracy Usher asked why SLF6 non-support for Python 3 matters since we build
everything from scratch anyway, and the answer was that we have not been
successful in building it on SLF6.

The EOL for Python 2 is Jan 2020.

Lynn Garren indicated that FIFE will begin pushing people to move to SLF7 at
the end of the summer, and EOL for SLF6 is Nov 2020.

The grid is already more than 50% SLF7, and there are very few sites that do
not have any SLF7 nodes available.

It was decided that there will be a test release of LArSoft with Genie v3.

Refactoring Services in the LArSoft Simulation
----------------------------------------------
by: Hans Wenzel

Worries were expressed by Alex Himmel about the memory usage of SimDriftedElectronClusters.

There was a discussion about storing parameters in GDML files versus storing them in FCL files.

Question was asked if ROOT can read the properties stored in the GDML files.  The answer is that ROOT does not have access to the properties supported GDML extensions, but can access everything else.

There was an externsive discussion comparing and contrasting the standard LArSoft drift simulation and the Wirecell drift simulation.

Lynn Garren feels there are a lot of action items before this can go into the release.

1. Rename ISCalculation
2. Material properties gdml vs fcl
3. fcl validation

Only item 1 needs to be done before release, and as this is trivial, the
request was approved for this week's release.

Breaking Reconstruction Workflows into TPC Units (for Icarus)
-------------------------------------------------------------
by: Tracy Usher

This change requires fcl file changes in the following packages:

larana - mcparticle <-> hit assns generation
larreco - cluster3d
lareventdisplay

The changes are all to replace a single InputTag as the input for an algorithm
with a fcl sequence of input tags as input to the algorithm.  This allows the
algorithm to be fed data one TPC as a time, or one cryostat at a time.

This change was approved for this week's release.