Interface Inference

The rules the inference draws its conclusion by, by IRI: a regularity of how the rolls were edited, such as that the same addition at the same place is not made twice independently, a default, such as that copies were not made from two exemplars, or a method, such as how the tolerance of a collation is found. The rules are stated once where they are published, and each inference names those it follows, so that a reader can see what a conclusion rests on besides its premises, and which conclusions rest on a rule.

References (by @id) to the beliefs from which the conclusion is drawn.

A belief the editor holds and publishes elsewhere under an IRI of its own, such as a bound from a catalogue of premises the editor keeps, is named by that IRI. Someone else's belief is not: it is adopted first, and the belief of the edition's own that the adoption concludes is named instead.

References (by @id) to what the inference worked on besides beliefs, such as the versions it compared, the analysis output it cites, or the document holding a premise named by IRI, in the state it was read in, since a belief published elsewhere may move after the edition has used it.