Physical atom definition and treatment of planar triple junctions #6
Description
There appears to be some ambiguity in the definition of physical atoms and the handling of triple junctions based on the choice of physical atom. The paper implies that the the physical atom for which bonded terms are removed is arbitrary, but this might be specific for the example in the paper, i.e. not true in general. In particular, the treatment appears to be broken when there is a physical atom in only one end state. In such situations, treating this as a non-physical atom produces modifications consistent with the factorisation requirements.
Another ambiguity is the treatment of planar triple junctions once the bonded terms for one physical atom have been removed. Should only remaining angles for the other physical atoms be modified, or should the junction now be fully treated as a dual junction?