Kwok-Bun Yue

A more general model for handling missing information in relational databases using a 3-valued logic

Codd proposed the use of two interpretations of nulls to handle missing information in relational databases that may lead to a 4-valued logic [Codd86, Codd87]. In a more general model, three interpretations of nulls are necessary [Roth, Zani]. Without simplification, this may lead to a 7-valued logic, which is too complicated to be adopted in relational databases. For such a model, there is no satisfactory simplification to a 4-valued logic. However, by making a straightforward simplification and using some proposed logical functions, a 3-valued logic can handle all three interpretations.

Priority ceiling protocol in Ada

The priority ceiling protocol (PCP) is an effective protocol for minimizing priority inversions in real-time scheduling. Priority inversion occw-s when a high priority task is blocked by a low priority task, such as at a shared semaphore or a protected operation. PCP guarantees the absence of chained priority inversion or deadlock. The ceiling locking (CL) priority on protected objects of Ada-95 also has these properties but has some limitations. For example, tasks cannot suspend themselves inside a protected operation. PCP has no such restrictions. Thus, PCP is more appropriate for some real-time applications. PCP has not been implemented in any language, · including Ada. A guideline for emulating PCP using Ada-83 exists but it lacks generality and flexibility. This paper discusses an implementation of PCP in Ada-95, the Ada-95 features that enable the implementation, the design of the implementation and some related issues.

An undergraduate course in concurrent programming using Ada

This paper describes a senior level course in concurrent programming using Ada. Unlike other similar courses in the subject area, it is not part of an operating systems course, nor is it tied to a particular hardware architecture. The course is software oriented and it discusses in depth a concurrent programming language, Ada, so that students are able to actually develop effective concurrent programs to solve problems in a wide range of applications. Ada is selected because of its popularity, superb portability, numerous hardware platforms, and rich concurrent constructs. Classical issues in concurrent programming are presented in the context of Ada. General issues in designing concurrent programming languages are elaborated using Ada, together with other concurrent programming languages such as CSP, Occam, and Linda. Finally, general principles of designing parallel programs are also discussed. Therefore, the course provides both the depth in a concurrent programming language for program development and the breadth in concurrent programming theory for insight. Using Ada throughout the course strengthens students expertise in Ada and provides an useful reference point for understanding concurrent programming theory. The course is heavily based on handouts, examples, homework and programming assignments. A rich set of instructional materials are available from the author.

Semaphores in Ada-94

The new protected type and requeue features in Ada-94 make possible the efficient implementations of many concurrent solutions, previously not possible in Ada-83. In this paper, the implementations of various semaphore systems in Ada-94 are presented. These systems are more general and powerful than the familiar binary and counting semaphores. As examples, the uses of the most general one, PVgeneral, in resource management and the synchronization of general mutual exclusion problems are discussed.

An Ada solution to the general mutual exclusion problem

Although some specific mutual exclusion problems have been studied extensively, automatic solutions to synchronize general mutual exclusion problems with arbitrary mutual exclusion constraints have not been fully explored. This paper discusses an Ada 83 solution that can be applied to any general mutual exclusion problem. This solution is based on strong binary semaphores. A generic package is used to generate the solution for an given mutual exclusion problem. An Ada program for the simulation of the classical Dining Philosophers Problem is presented as an example to demostrate how the solution can be used.

Dining philosophers revisited, again

This paper describes a problem in the solution of the dining philosophers problem by Gingras [2] that makes it inefficient, instead of the claimed maximal efficiency. A correct implementation is presented. Even then, the solution has other undesirable characteristics and is still not maximally efficiency. Depending on the definition of efficiency, it may not be possible to attain maximal efficiency and be starvation-free at the same time. A better and simpler solution for general mutual exclusion problems, in which the dining philosophers problem is a special case, is presented. This solution can become symmetric if appropriate data structures are used.

Teaching a graduate expert systems course

Because expert systems technology is relatively new, there are only a few papers describing a graduate course on the subject. Most of the described courses were either seminar oriented or centered around a student project, but did not have a complete coverage of the major topics. This paper describes the teaching of a graduate expert system course that has a relatively complete coverage of the subject area. Other additional features of the course include incremental development of the student projects, emphasis on evaluations of tools and projects, as well as using more conventional problems for strengthening expert system shell programming techniques.

An Optimal Algorithm for Reducing Edge-Solvable Mutual Exclusion Graphs

A mutual exclusion graph can be used to model a graphical mutual exclusion problem where a concurrent process is represented by a vertex and a mutual exclusion constraint between two processes is represented by an edge joining the two corresponding vertices. The first step of an algorithm for generating very efficient semaphore-based edge-solvable synchronization solutions for a class of graphical mutual exclusion problems calls for reducing the mutual exclusion graph to an optimal final graph with a single composite vertex representing them. The process of reduction involves the identification of similar vertices (with identical neighbors except, perhaps, the vertices themselves) and the replacement of the similar vertices by a composite vertex