I am interested in tracking down any paper documenting the application of
circumscription (cf. McCarthy, Lifschitz) to

1. applications other than the minimisation of ab predicates ?  In
        this class, I do not include papers showing the relationship
        between circumscription and, for example, logic programs but
        do include applications such as formalising configuration
        tasks (which typically minimise the sets of components being
        assembled in a configuration).

2. actual applications of circumscription to diagnosis,
        showing practical application in some domain ? While
        Reiter’s work on diagnosis from 1st principles certainly
        recognises the relationship between diagnoses being minimal
        sets of abnormalities, I am not aware of any papers that
        document the application of a circumscription policy to a
        set of formulae and subsequent computation of the diagnoses.

As 1. suggests, I am considering the feasability of describing a class of
configuration problems in terms of circumscription policies applied to a
first order theory and am interested in whether any work has been done in
applying circumscription in domains other than directly to default
reasoning.  Applying circumscription policies provides a conceptually nice
way of stating predicate minimisation (such as that a minimum number of
components satisfying the constraints of a design problem is often a
desirable configuration).  Obviously, the irreducability to first-order of
general first-order theories is of some concern but potentially controlled
if the class of admissable formulae are suitably constrained.

sincerely,

Nirad.
—
Nirad Sharma
Computer Science, University of Queensland. 4072.  Australia