Some discussion about the options for HepMC.  

There is no concern with the PdfInfo class, but we think that kf1 and kf2 should be renamed.  See Torbjorn's talk in this session for better names.

There was some surprise expressed at the use of the FourVector != within HepMC.

We agreed to add generated mass as a double in GenParticle.  In principle, experiments can choose not to save the generated mass.  However, it is strongly advised that, if storage space is an issue, experiments should save the 3 momentum and the generated mass.  The 4 vector can be safely reconstructed from 3 momentum and mass, but mass cannot always be safely reconstructed from the 4 momentum.