Copyright:		© by the author
Author:		Hassan Aït-Kaci
Version:		Last modified on Wed Aug 07 15:31:11 2019 by hak

It adapts previous versions consisting of a bare-bone syntax and adds a syntactic operational semantics consisting of keeping order-sorted feature (OSF) terms in a well-defined normal form — also known as ψ-terms — implementing basic operations for them. It first focuses on defining a multiple-inheritance concept hierarchy of sort symbols. It implements the machinary for compiling these sorts into binary codes to optimize lattice operations on them for the inheritance ordering on ψ-terms.

The latter provide basic KR reasoning functionality with ψ-term unification (returning the greatest lower bound of two terms) implementing conjunction for deductive querying, and generalization (least upper bound of two terms) implementing non-modular (and therefore non-distributive) disjunction for inductive learning, as well as all the necessary inferential normalizations. Supported in this version:

• built-in constants (paired with their types);
• object and set types;
• "Strict/Permissive" feature mode;
• classification by transitive closure, feature propagation and consistency checking, and feature-aware normalization;
• RDF XML serialization consistent with the one described in the basic specification.

• consistent syntax for TBox, ABox, and QBox processing;
• modular URI-based TBox, ABox, and QBox processing;
• a SPARQL interface;
• Horn and functional rules, residuation, lazy inheritance, DataLog interface etc.

Thus, this version of HOOT is meant to be the last of the initial prototype for declaring a sort lattice, implementing binary encoding of sort ordering using bit vectors, evaluation of sort expressions, OSF unification and generalization with featured sorts, values, and sets, and reading and pretty-printing ψ-terms. It also offers interactive pragmas for exploring a sort hierarchy. We hope to build on this basis to add all these further emendments and extensions in coming HOOT versions.