Present: Vito Di Benedetto, Giuseppe Cerati, Christian Farnese, Lynn Garren, Krzysztof Genser, Kyle Knoepfel, Robert Kutschke, Marc Paterno, Gianluca Petrillo, Brian Rebel, Saba Sehrish, Erica Snider, Filippo Varanini, Hans Wenzel, Tingjun Yang
Remote: "BNL", Lorena Escudero, Alex Himmel, Yun-Tse Tsai
Releases and project report [Erica Snider]
-
new art 2.6 supports OSX 10.12 (Sierra) and Ubuntu 16.04 LTS
-
LArSoft does not support those yet, it will when we switch to art 2.6
-
[Lynn Garren] there is a set of non-trivial changes presented in a previous meeting by
-
[Kryzsztof Genser] provisions to have both Geant4 versions supported during validation time? [A] We use to allow validation on a release candidate, but no support for multiple
-
we'll pass to using Zoom instead of Readytalk starting from the next week
Conclusion
-
the upgrade plan to art 2.6 has been accepted with no objections
-
LArSoft we update to Geant 10.3p01
-
a load of new features: event mixing for runs and subruns, extended association lookup, and
-
art tools, plug ins that allow module behaviour
-
each tool instance will be "local" to the module
Shower reconstruction/changes to recob::Shower [Yun-Tse Tsai]
-
[Brian Rebel]
-1 is not a number stupid enough for an invalid (default) value; std::numeric_limits<float>::lowest() may be a better choice
-
[Robert Kutschke] accessors might be better throw if the value they access is not valid [A] that was thought to be put in a facade, leaving the base data product free of overhead
-
[Brian Rebel] the underscore-separarted naming style is inconsistent with the rest of large part of some of LArSoft
-
[Erica Snider] being geometric properties, we should agree on their definition
-
[Yun-Tse Tsai] for different reconstruction algorithms, there might be different recipes
-
[Robert Kutschke] if there is a lot of action on the opening angle, it may be worth having it as a data product
-
[Thomas Junk] shower opening angle might be not conical [A] it may be a good future addition
-
requesting to merge the feature branches into next release
Conclusions
-
we will discuss with the experts whether using
std::numeric_limits is safe for comparisons and consistent storage
-
we will add the restyling of class names to the list of actions for the future
-
the merge is approved
Radiological generator and geometry [Thomas Junk]
-
radioactive decays of Argon are simulated in the whole volume, which is not realistic (and hurts when happening "inside the photodetector"
-
a new mandatory configuration parameter is specifying which materials are
Conclusions
-
the old way to simulate argon decay should go away; Thomas Junk will follow up with Hans Wenzel
-
no objection to the breaking change
Optical simulation changes for SBND [Diego Garcia-Gamez]
Speakers will present at a later date.
There are minutes attached to this event.
Show them.