Gregor Meyer

Data mining standards initiatives

Lacking standards for statistical and data mining models, applications cannot leverage the benefits of data mining.

Run-time type computations in the Warren Abstract Machine

Abstract The type concept of the logic programming language PROTOS-L supports sorts, subsort relationships, and parametric polymorphism. Due to the order-sortedness, types are also present at run time, replacing parts of the deduction process required in an unsorted version by efficient type computations. Together with the polymorphism, most of the flexibility of untyped logic programming carries over the order-sorted approach. The operational semantics of PROTOS-L is based on polymorphic order-sorted resolution. Starting from an abstract specification, we show how this operational semantics can be implemented efficiently by an extension of the Warren Abstract machine, and give a detailed description of all instructions and low-level procedures responsible for type handling. Since the extension leaves the WAMs AND/OR structure unchanged, it allows for all WAM optimizations like last call optimization, environment trimming, etc. Moreover, the extension is orthogonal in the sense that any program part not exploiting the facilities of computing with subtypes is executed with almost the same efficiency as on the original WAM.

A pragmatic type concept for Prolog supporting polymorphism, subtyping, and meta-programming

Abstract Approaches to typing logic programs often exclude various features of Standard Prolog. The system “Typical for annotated Prolog” (TaP) is a pragmatic approach to type checking programs written in Prolog without restricting the scope of the language. TaP checks Prolog programs that are extended with type declarations that support parametric polymorphism and subtyping. The purpose of this paper is to present an approach that extends Typical by meta-types for handling Prolog meta-programming techniques.

Using Types as Approximations for Type Checking Prolog Programs

Subtyping tends to undermine the effects of parametric polymorphism as far as the static detection of type errors is concerned. Starting with this observation we present a new approach for type checking logic programs to overcome these difficulties. The two basic ideas are, first, to interpret a predicate type declaration as an approximation for the success set of the predicate. Second, declarations are extended with type constraints such that they can be more refined than in other conventional type systems. The type system has been implemented in a system called Typical which provides a type checker for Standard Prolog enriched with type annotations.

A Brief Description of the PROTOS-L System

The PROTOS-L system is an outcome of a research effort within the EUREKA Project PROTOS (Logic Programming Tools for Building Expert Systems, EU 56) aimed at overcoming some of the shortcomings of Prolog. It consists of a logic programming language including various advanced features together with a state-of-the-art implementation approach based on an abstract machine.

Aspects of Coupling Logic Programming and Databases

Emphasizing the pragmatic aspects of a close coupling of logic programming and databases, six coupling levels are distinguished, starting from a simple embedded SQL interface to using a deductive database component. As a concrete example, the PROTOS-L system is presented where several of these levels have been realized. A special database inference component extends the capabilities of a relational database to support Datalog plus additional SQL functionality within PROTOS-L. Furthermore, there is an integrated access to the non-standard database system LILOG-DB which allows to use complex structured objects not only in the logic programming language but also in the database.*