HepMC Technical meeting minutes - 15/06/2011
Developments related to ROOT I/O:
- in the solution proposed by Lynn, GenEvent would hold a vector of pointers to GenParticle objects and a vector of pointers to GenVertex objects
- the navigation between particles and vertices would most likely be done using the indices of the entries in that vector
- in order for the indices to stay unchanged, if particles/vertices are removed from the event, the pointers should be replaced by null pointers (and not removed from the vector)
- the map of barcodes to particles would be transient or would be encapsulated
- Lynn is implementing the prototype, it should be available for testing soon
- other potential solution would be to investigate again the use of barcodes for the navigation between particles and vertices
CMAKE:
- an alternative build system based on CMAKE will be implemented for 2.06 series; this should solve the problem of building on Windows
- problems related to libtool should also be solved once we move to CMAKE
Examples:
- will add new examples based on C++ generators (Pythia8, Herwig++); the code with the new example may be provided by the GENSER team
Condensed ASCII I/O:
- Leif Lonnblad reported from the Les Houches meeting where an idea was raised to implement ASCII I/O that would produce HepMC files largely reduced in size. The current ASCII writer produce very big files, difficult to work with, but depending on your use case, much of this information may be unnecessary. For instance, proposals include only keeping particles with status 1 and 2, reducing precision, saving a 3 momentum and mass instead of 4 momentum. The new writer would allow to produce an 'intelligently compressed' file, using various switches. Lynn has met with Leif and Gavin and will be preparing a prototype. This will be initially implemented as an additional module to 2.06. (and eventually integrated into 'core' HepMC)
End of support for 2.03 series:
- It was proposed to the experiments to stop the support of the 2.03 series by the end of 2011. This would be the deadline for the migration to 2.06 series which then would become the official production version used by all the LHC experiments.
