Our model for IUI assignment errors comes from Dealing with Mistakes in a Referent Tracking System (Ceusters, 2007).

We have currently enumerated the specified errors, but additional error types are mentioned in the passage:

... entries expressed through PtoP-tuples relate particulars to each other and may
involve many further sorts of mistakes (P1, P2, …), depending on whether one or both IDs
involve an A-type of mistake, and whether the relationship in question holds.

After considering the "plus" types, it seems that only these three would "convert" to a particular error (or, an error involving a PtoP template):

P1("The relationship between the particular referred to by the IUI and "
+ "the particular in question does not hold during the stated time "
+ "period"),
P2("The UUI for the particular does not refer to the intended particular or "
+ "it refers to no particular at all"),
P3("There is an A1 error in the corresponding A-tuple; the PtoP-tuple is "
+ "nonsensical"),

After discussion of the "minus" types, some errors are mentioned as being always related:

Whereas mistakes of omission may occur independently of other mistakes, some mistakes of type A and type U will
automatically bring in their wake mistakes of other types: Thus for example, mistakes of type U6 and U9 will be
automatically associated with a mistake of type U–1.

It would seem our software should have constraints in this regard.

TODO: Define "propagate the resolution"

Related work for ontologies in: A Realism-Based Approach to the Evolution of Biomedical Ontologies (Ceusters and Smith, 2006).

Browser not supported