OWLlink provides a declarative interface for – among other things – asserting OWL axioms to a Knowledge Base (KB) of an OWL reasoner. This OWLlink extension adds the ability to retract previously asserted axioms from KBs.
The OWLlink Working Group seeks public feedback on this Working Group Recommendation. Please send your comments to public-comments@owllink.org or post to the public discussion forum at http://www.owllink.org/forum/.
This extension is being published as one of a set of 4 documents:
The OWLlink core interface [OWLlink Structural Specification] relies on a certain kind of incremental batch-oriented reasoning procedure. After submission of a set of axioms a client can pose some queries, add further axioms, query again, and so on. Deletion of axioms is ony possible by releasing a KB and re-submission of axioms. This document describes the retraction extension of OWLlink which allows to retract previously told KB fragments on axiom level.
This extension is based on the same preliminaries than the OWLlink core.
The Retract request is the inverse command to Tell (cf. Section ‘4 Tells’ [OWLlink Structural Specification]) and takes a set of OWL 2 axioms (cf. Section ‘9 Axioms’ [OWL 2 Structural Specification]) and removes them from the respective KB as shown in Figure 1. The removal must be sensitive to the rules of structural equivalence of OWL 2. That is, an axiom can only be removed from a KB if and only if the KB contains an axiom which is structural equivalent to the one given in the retraction request. A server must not respond with an error if a KB does not contain a structurally equivalent axiom but is free to provide a warning. If all axioms of a retraction request are successfully removed from the KB, the server has to respond with an OK response.
Retraction Request (the Tell request is rendered semi-transparently to show the symmetry)