Viotti and Vukolic surveyed over 50 consistency semantics (or models) to understand how all of these semantics relate to each other. They created a diagram showing the partial ordering of these consistency semantics, which range from weak consistency to linearizability (strong consistency). In one section, they describe their semantics and terminology in order to coalesce other consistency semantics into the same semantic notation, and then in a following section they provide their analysis of the consistency semantics and compare their strengths in order to determine the actual partial ordering and an additional, loosely-defined clustering of consistency semantics into families.
Without context, I thought the partial ordering of consistency semantics was a total ordering. Knowing now that it’s just a partial ordering makes me realize that relationships between consistency semantics that don’t have an edge probably isn’t well understood. I have been interested in understanding a total ordering of consistency semantics in order to understand if some semantics are totally included in others. I suppose that a partial ordering does just this, and is just unable to order semantics that are sufficiently different, in the sense that first order logic is sufficiently different from propositional logic.