We should specify that the internal data structure used is bi-partite FG (in the developer's documentation)

[bvdmitri]

It might lead to a confusion, especially when we will have the visualisation in a form of FFG. The internal data structures do not use FFG representation for efficiency. But in many places we mention FFG and people get confused, I think we should clarify that. (Also in RxInfer).

On the surface, I think we should just say factor graphs, instead of Forney-style factor graphs. For users it doesn't really change much since we can visualize an arbitrary FG as an FFG, but for the developers who actively work with the code it may reduce the confusion.