Stakeholders / Partner and Project (27 May 2014)
OK to add leading 0 for major version number
Proposal for Documentation Requirements - in addition to doxygen which is the reference manual. The:
in Examples: Link to ficl files that are used and work.
Add Credits: author, owner, contributors, papers that are references.
Add at top "Documentation of order a page in length"
Will need to export built PDFs to some common page. Action item to discuss where things should be placed.. Should LArSoft have a docdb - so documents can have metadata?
Add milestone for NIM article Oct 1 2014
Agreement to have merge back after discussion at each librarians meeting.
Trigger is very detector specific and this affects the simulation code - so there is a stress to know how to do this in ubdev which is in LarSim and will need to be merged into the dev release. Potential impact on the architecture of LArSoft. Each time there is a need for a branch could stimulate a discussion of refactoring the main classes rather than make a branch.
- Erica will look at the current LarSim:ubdev code and discuss the issues. some of it is optical. Optical is a big issue in general.
- Erica will schedule a meeting to continue discussion of principles/need/impact of having experiment branches - include Kazu and Seligman. Could invite Brett.
Make it more clear that FIFE and LArSoft workshops are complementary.
LBNE:
Brett - get jenkins running at BNL. Let LarSoft know and we will send jobs to BNL.
LArSoft Project Will post any documents we get from Patrick and distribute the links.
MicroBoone:
Kazu and Timeservice. An art service that will change timescales among art data products: Geant, Trigger, Beam. Will not impact other experiments. Also talking about time being measured/coordinate system/unit wrt experiment time e.g. Global T0. Will discuss this at the next Librarians meeting. Need to extract/discuss impact on other users.
LArLite: Wes pushing that algorithms should be framework independent. Erica and Wes have spoken.
Erica's principle is that algorithms should have very clean interfaces with only the algorithm internally. Issues identified: 1) not allowed to access art service including geometry; this requires a Change in interface which is not thought to be too bad; 2) Agreemeent to have LarLite receive configuration via FICL.
Brett: One of the work items BNL will work on for LBNE is a framework independent interface to the various services such as Geometry, event data etc. Will need specific funding from DOE for this and it is not known if this will happen yet. Erica believes this can be relatively straightforward work. Will simplify the testing/validation framework as can think of unit tests. Erica continuing discussions on the topic. Herb: will need to identify service labor for someone to go through a change all the interfaces centrally. Erica: not all the services can be handled in the same way. Need to have a careful survey of the work involved - first deliverable of current proposal. Work will need to be done in close collaboration with the art team.
Workshop
Have someone from Panagiotis talk about CI systems - e.g. best practices. Get topic.
FIFE - CI System - Glen
Worch- Brett
Unit Test Writing - ask Marc Paterno. Larsoft or FIFE workshop? Survey of Boost Unit tests. Better for the LarSoft workshop. Developer types.
Additional topics:
Opitcal - Ryan Wasserman (LBNE), Eric Church, Ben Jones (uboone).
LArPandoraAlgorithms - Andrew Blake :
There are minutes attached to this event.
Show them.