African Journals Online
South African Computer Journal / Die Suid-Afrikaanse Rekenaartydskrif

Issue 26, November, 2000
Abstracts

A comparison of bisimulation-based semantic equivalences for noninterleaving behaviour over CCS processes

Galpin, V.

Abstract: A number of extensions to the process algebra CCS (Calculus of Communicating Systems) have been proposed to deal with noninterleaving behaviour such as location and causality. The aim of the paper is to use existing and new comparison results to provide a hierarchy of these semantic equivalences over pure finite CCS terms. It is not possible to include some extensions in this hierarchy and the reasons for the exclusion are given.

Algebraic results for structured operational semantics

Galpin, V.

Abstract: This paper presents algebraic results that are important for the extended tyft/tyxt format [12, 13] which can be used to describe many different process algebras. This format is based on a many-sorted signature which permits both processes and labels to be treated syntactically. Existing results for this format permit the comparison of process algebra semantic equivalences by forming the sum of two transition system specifications and imposing certain conditions. The results presented in this paper involve the summing of congruences that model the actual process algebra labels, and determine under what conditions these congruences have important properties such as compatibility and conservativity. The aim of this paper is to show that the notion of sort-similarity on the sum of signatures is sufficient for the sum of the congruences induced by each label algebra to be the same as the congruence induced by the summed label algebras. Additionally, sort-similarity is sufficient for compatibility and conservativity when summing. Finally, conditions on the label algebra are given that ensure compatibility.

Algorithms for the shortest exact common superstring problem

Elloumi, M

Abstract: The Shortest Exact Common Superstring (SECS) problem is: Given a set of strings f = fw1; w2; : : :;wng, where no wi is an exact substring of wj ,i6 = j, find a shortest string Se such that every string of f is an exact substring of Se. When the number of the strings n > 2, the SECS problem becomes NP-complete. In this paper, we present an exact SECS algorithm and a greedy approximation one. Our greedy algorithm is a 1/2-approximation for the SECS problem. Our exact SECS algorithm and our greedy one are, respectively, of complexities O(n n ) and O(n 2 _ (l +log(n))) in computing time, where n is the number of the strings and l is the maximum length of a string. Our SECS algorithms are based on the computation of the Length of the Exact Longest Overlap (LELO).

An expert system algorithm for the identifcation problem

De Kock, G de V

Abstract: An algorithm is proposed for an expert system to solve the general identifcation problem which is formally defined. The algorithm follows a statistical approach and is based on the work of the author in [2] and Vogts in [11]. A measure is defined to express the similarity between an observed specimen to be identified and any one of the possible classes. It is based on a similarity function of the common property values. This approach solves the inaccuracies of the linearized approach given in [2]. Using a measure of clustering, a method to select the `best' next property to be established during the execution of the algorithm is proposed. A prototype of the algorithm is tested ona small knowledge base to identify South African cycads, similar to that given in [2]. The comparison with the results in [2]shows that the proposed algorithm is a significant improvement.

A formal perspective to specification of transaction systems

Ehikioya, S.A.Barker, K.

Abstract: This paper describes the relationship and relevance of predicate logic in the specification of transaction system protocols.We present a formal specification technique based on concurrent execution that provides automatic communication and synchronization mechanisms between concurrent processes. This framework is not biased towards specific types of trans-actions.It integrates temporal behaviour of individual transactions with the dependencies that can arise when accessing shareable data. Our approach allows the specification of transaction constraints and transaction dependencies in a declarative fashion. This provides flexibility in their implementation and permits different techniques to enforce the constraints independent of the application program. This paper explores the issues of correctness, reliability, and recovery of data distributed across the enterprise; describes techniques for guaranteeing and enhancing correctness; and procedures for recovering transactions and data in case of any transaction failures to guarantee a high degree of system dependability.

A multiple perspectives evaluation of the factors affecting software development productivity in an outsourced IT project

Petkova, O.Roode, J.D.Petkov, D.

Abstract: The paper provides a model for examining the perceptions of the users and the outsourcing providers about what influences software development productivity. It is illustrated through a case study. The model is using a multicriteria decision analysis approach - the Analytic Hierarchy Process. This type of analysis of perceptions of the different stakeholders in an outsourcing relationship can be applied to other organisational environments to help improve the management of outsourced IT activities.

An incremental construction algorithm for Venn diagrams

Eloff, J.van Zijl, L.

Abstract: The drawing of Venn diagrams have traditionally served as an aid to analyze the characteristics exhibited by the diagram and its underlying Venn graph. Little attention has hitherto been given to the presentation, construction and way in which a Venn diagram conveys information. We present a new algorithm for the construction and layout of a Venn diagram based on an incremental technique. The algorithm constructs the Venn diagram by incrementally deriving new intersections from previously placed ones on a grid. We also propose criteria which can be applied to analyze a Venn diagram's drawing and layout. Our algorithm aims to generate an aesthetically pleasing layout by satisfying the proposed criteria.

An XML based approach to enforcing history-based separation of duty policies in heterogeneous workflow environments

Papenfus, C.Botha, R.

Abstract: In the computing world a new technology occasionally comes along, promising to make dramatic changes to the way computing tasks are performed. The Extensible Markup Language (XML) has been heralded as one such technology. XML promises to provide a universal metadata mechanism for defining, understanding and interchanging information between possibly heterogeneous systems. This paper exploits this powerful promise of XML by examining how it can be used to enforce history-based separation of duty policies across heterogeneous workflow environments. A very brief overview of separation of duty policies is provided, whereafter the need for history-based separation of duty is motivated through an extensive case study. A solution based on XML baggage is proposed and it is shown how the solution would operate in the context of the case study.

A proposed framework for evaluating generic enterprise models

Van Belle, J-P.Price, B.

Abstract: This paper investigates a number of ways in which enterprise models can be compared and proposes a framework for assessing the quality of generic enterprise models. The paper starts with an overview of various qualitative and quantitative yardsticks, taken from related disciplines, which can be used to evaluate generic enterprise models. It then uses some of these to develop a suitable framework for assessing or comparing the quality of enterprise models. The resultant framework is essentially an extension of Bart-Jan Hommes' framework for analysing the quality of a business modelling technique, to include some criteria suggested for evaluating the genericity of enterprise reference architectures. To illustrate how the proposed framework can be operationalised, it is then applied to two well-known “general enterprise models”: the reference models underlying two ERP solutions namely those of Baan and SAP AG.

Computerised assessment of the appearance of abraded fabric

Fazekas, Z.Szalai, L.

Abstract: As a continuation of our earlier research on automatic visual measurement methods of fabric quality features, here we present an automatic visual approach towards measuring another fabric quality feature, the feature being the durability of the fabric's visual appearance (DVA). This is a novel fabric quality feature, though a closely related quality feature is sometimes used for certain types of fabrics. DVA complements the presently frequently used fabric quality feature -namely abrasion resistance - without imposing prohibitive extra procedures into the standardised fabric assessment. The DVA describes the way the fabric's appearance changes during its mechanical lifetime. Effectively, it furnishes information on the portion of the fabric's mechanical lifetime for which the fabric remains visually (aesthetically) stable. In our paper we explain why the DVA is a useful fabric quality feature for the various players in the fabric trade and show how the measurement of a fabric's DVA relates to the measurement of abrasion resistance. We also present a possible implementation of the suggested assessment approach. The DVA of a fabric sample is calculated from images taken at different stages of abrasion testing. n the presented implementation certain brightness, colour and morphological features are calculated for each of these images and the acquired feature values are used to calculate the fabric's DVA.

Cooperative learning in neural networks using particle swarm optimisers

Van den Bergh, F.Engelbrecht, A.P.

Abstract: This paper presents a method to employ particle swarms optimizers in a cooperative configuration. This is achieved by splitting the input vector into several sub-vectors, each which is optimized cooperatively in its own swarm. The application of this technique to neural network training is investigated, with promising results.

Diversity and effectiveness in information systems project teams

Smith, D.C.Van Jaarsveld, QNeethling, C.Andrews, R.

Abstract: This study investigated IS team effectiveness in relation to team diversity.Team diversity in the context of this research was based on Belbin Rs Team Role Theory. 12 student project teams R team role diversity were evaluated using 2 commonly used Belbin questionnaires. A third questionnaire, based on the Francis & Young model, was used to evaluate overall team performance levels. An analysis of the results dentified a very strong relationship between team role diversity and team effectiveness in IS project teams. Thus, the greater the role diversity of the team, the better the team performance. These conclusions support the findings of other more general research studies and can assist those selecting IS project teams who want to improve overall team effectiveness.

Evaluating the BPR effect of a SAP R/3 implementation in a manufacturing environment

Calitz, A.P.Calitz, M.B.

Abstract: The introduction of ERP software in organisations changes business processes and procedures. ERP implementations have therefore been compared to BPR projects, which focus on streamlining business processes. Critical success factors have been identified for both ERP and BPR implementations. In this study it is shown that the implementation of ERP software has a business process reengineering effect on the business and that the business should therefore also focus on BPR success factors. Various attributes are associated with reengineered processes, such as faster throughput, empowerment of employees and elimination of unnecessary tasks. The implementation of an ERP package, SAP R/3, at a textile manufacturing company was effectively a reengineering project as is proven in this study, where the resulting processes are measured against attributes of reengineered processes.

Feedback in human-computer interaction — characteristics and recommendations

Renaud, K.Cooper, R.

Abstract: The need for feedback is widely accepted, and the role of feedback in making things easier for the end-user is also no longer disputed. However, the nature of the feedback and the manner of this provision is by no means agreed upon. Feedback is traditionally considered to be a communication of immediate system state to the end-user. This paper will argue the need for extending the concept to encompass the provision of archival feedback — information about past activity. We also argue the need for graphical, rather than textual, feedback and provide a list of desirable feedback features which should be provided by an application. The paper ends off by examining the means for this provision.

Investigating the use of multiresolution image processing and artificial neural networks for computer-aided diagnosis within a telemedicine network

Aydn Alaylolu, B.Aghdasi, F

Abstract: In image processing and pattern recognition, the usefulness of extracted image features and classifiers are assessed according to their accuracy in classifying new data, such as assigning a diagnostic decision to digitised medical images. In such applications, predictive classifiers are a valuable tool only if they increase the number of true detections while minimising the occurrence of false predictions. A computer-automated detection scheme, making use of an artificial neural network classifier, with input feature vectors containing spatial, spectral and multiscale image attributes, is investigated here with an objective to improve the efficiency of the medical diagnostic processes. A wavelet-based image enhancement technique is also implemented, with the aim of improving the computer-automated detection performance. The incorporation of the CAD workstations into a telemedicine service network and biomedical informatics systems to meet the challenges of the new millennium is described.

Knowledge management technology: examination of information diverse repositories

Handzic, M.Parkin, P.

Abstract: This paper reports results of an empirical examination of the contribution of information diverse repositories in enhancing individual knowledge and performance in a judgemental decision making context. A laboratory experiment was conducted using 32 graduate students as voluntary subjects. Performance of actual subjects was compared with that of their nominal naive and optimal counterparts. Results indicate that actual subjects performed better than naive, but worse than optimalnominals irrespective of the level of information diversity present in their available repositories. The results also indicatethat subjects tended to perform significantly worse when faced with more diverse information. It can be concluded from the results that, in general, knowledge workers may have difficulties in turning available task information into knowledgand translating it to task performance. This also suggests that such workers may potentially benefit from a number of other knowledge management initiatives that would enhance their understanding of the existence and the form of relationships among diverse information.

Motivating and recruiting intending IS professionals: a study Of What attracts IS students to prospective employment

Turner, R.Lowry, G.

Abstract: This paper is the second study in a research program aimed at achieving a better fit between university courses and professional practise of information systems. The paper reports the results of a survey of student attitudes towards incentives and conditions of employment that may be useful to employers who wish to attract and retain scarce and talented information technology professionals. Students were asked to rate the importance of eleven conditions and incentives and to identify and rank the five incentive variables that they considered most important to them in their decision to accept and remain in work in a given organisation. The findings suggest that IS students are most attracted by a friendly work environment, supportive superiors, and perceived promotional opportunities, with economic incentives such as salary and fringe benefits rated as rather less important. Implications for employers who are dealing with an ongoing global shortage of information technology staff are discussed.

On the validation and legitimisation of an information system-some theoretical considerations and a case study

Petkova, O.Petkov, D.

Abstract: This paper provides a discussion on the validation and legitimisation of information systems in general. The theory is illustrated on the case of a project about the computerisation of the management of the research function at an university. The paper aims to show that traditional aspects of model validity and legitimisation in Operational Research can be applicable to the field of Information Systems (IS). However the validation considerations in this case sometimes need considerable modification since a more appropriate way of viewing complex IS projects is through an interpretivist viewpoint. Some extensions on the notion of validation for soft systems are provided for that purpose. The importance of both validation and legitimisation in Information Systems is illustrated on a case study about a failed software project.

Placing axial lines in urban grids

Sanders, I.D.

Abstract: In earlier papers the orthogonal axial line placement problem for orthogonal rectangles and the non-orthogonal axial line placement problem in orthogonal rectangles were presented. These problems were shown to be NP-Complete by transformations from the vertex cover problem for planar graphs. The general axial line placement problem – placing axial lines to cross the adjacencies between adjacent convex polygons - is a more general case of the problem of placing non-orthogonal axial lines in orthogonal rectangles and the NP-Completeness proof can be extended to this problem as well. In this paper the axial line placement problem and the related problem of generating the convex map for polygons with holes which can be represented as urban grids are considered. The paper shows that for these arrangements of polygons the solutions can be found in polynomial time. It is also conjectured that some more general configurations could also be solved in polynomial time.

Scheduling on a multiprocessor-based telephone switch

Liu, M.Majumdar, S.Streibel, D.Carroll, B.

Abstract: This research is concerned with reducing memory contention through process scheduling in the call control module of a scalable shared memory multiprocessor-based telephone switch. Multiple processes may contend for shared variables stored in the shared memory of the multiprocessor system. A lock-rollback mechanism is used on the system to serialize access to shared data and preserve data integrity. A process locks a memory line before accessing it. If the line is already locked, either the requesting process or the process currently holding the memory line is rolled back. The rolled back process restarts from the beginning at a future point in time. Rollbacks expend work to preserve data integrity. Effective control of memory contention is thus required for achieving high system capacity and scalability. Based on trace-driven simulation we have investigated the performance of three different scheduling approaches. These include two “pure” and a hybrid strategy that combines the pure strategies into a single scheduling approach. The results of the research demonstrate that the strategy used for scheduling can have a significant impact on performance. The strengths of the pure strategies are described and the tunability of a hybrid strategy for handling a wide range of workloads corresponding to a variety of different products is discussed.

Small group collaboration and presence in a virtual environment

Casanueva, J.Blake, E.

Abstract: Presence in Collaborative Virtual Environments (CVEs) can be classified into personal presence and co-presence. Personal presence is having a feeling of “being there” in the CVE oneself. Co-presence is having a feeling that one is in the same place as the other participants, and that one is collaborating with real people. In this paper we describe an experiment to investigate the effects that small group collaboration and interaction has on personal presence and specially co-presence in a CVE. We hypothesise that collaboration and interaction enhances co-presence in a CVE. We found that there was a large difference in co-presence between two CVEs which produced different levels of collaboration and interaction. This supports our hypotheses that just having virtual representations of others is not sufficient to create a high sense of co-presence, and that one needs collaboration and interaction in order to enhance co-presence in a CVE. We also found that. We measured personal presence subjectively, using a questionnaire developed by Slater et al. We have developed a co-presence questionnaire which assesses the levels of co-presence subjectively. We have also developed a collaboration questionnaire which measures group collaboration subjectively, as well as the degree of enjoyment and comfort with others in the group.

Structuring knowledge acquisition in software development projects

Trimble, J.A.

Abstract: The goal of this research is to provide knowledge acquisition approaches that will contribute to reducing the risk involved in software development. Structured knowledge acquisition techniques can increase the rate of obtaining and developing knowledge, and the reliability of the knowledge acquired. This project builds on knowledge acquisition research from the expert systems and project management sub-disciplines. Knowledge representation is examined as a component of the knowledge acquisition process. This paper is intended to provide background material and guidelines that will assist in structured interaction with stakeholders in the software development process.

Using the Taylor expansion of multilayer feedforward neural networks

Engelbrecht, A.P.

Abstract: The Taylor series expansion of continuous functions has shown - in many fields - to be an extremely powerful tool to study the characteristics of such functions. This paper illustrates the power of the Taylor series expansion of multilayer feedforward neural networks. The paper shows how these expansions can be used to investigate positions of decision boundaries, to develop active learning strategies and to perform architecture selection.

Website literacy requirements: readability and occulacy measurement development

Licker, P.Bailey, D.Scott, A.Stamper, B.

Abstract: This paper explores the concept of “literacy requirements” for the web and in doing so examines a number of different ideas about literacy. The concept of readability is one important component of effective literacy with regard to websites. The research reported on here applies several existing and newly derived measures of readability to a selection of websites. Subsequent analysis is used to draw conclusions about the validity and applicability of these measures to determining website readability. A number of literacy measures were found to be appropriate for determining the readability of web page content. Two specific measures are the most significant in relationship to actual reading and navigating through online content. Further research into web literacy is proposed.

A graphical environment for the facilitation of logic-based security protocol analysis

Saul, E.Hutchison, A.C.M.

Abstract: The development of cryptographic logics to analyze security protocols has provided one technique for ensuring the correctness of security protocols. However, it is commonly acknowledged that analysis using a modal logic such as GNY tends to be inaccessible and obscure for the uninitiated. In this paper we describe a graphical tree-based specification environment which can be used to easily construct GNY statements using contextualized pop-up menus. The interface which we describe helps to move logic-based analysis out of the world of academia and into the mainstream market.

Analysing algorithms using computed values

Mueller, C.

Abstract: The purpose of this short paper is to give some insights as to how computed values can aid algorithm design. A simple algorithm of the finding average is expressed in terms computed values illustrating how one can express an algorithm in a paradigm independent form. This same problem is used to show how a refinement process enables an algorithm to be developed from a specification. The refinement process enables alternatives to be explored at each step of the design. By looking at the number of computed values and their dependencies provides some simple tools for analysing the choices, this enables decisions to be made as to which alternative to select. Some simple theory is developed for describing the computed values. The advantage of the theory is that it is based on commonly used mathematical notation and restricted form predicate logic. Hence anyone with some basic mathematical background is able to understand the semantics of the theory.

An object oriented approach to parser generation in C++

Cosgrave, L.Power, J.Waldron, J.

Abstract: In this paper we describe the design and implementation of a system for representing context-free grammars in C ++ . The system allows for grammar representation at the object level, providing enhanced modularity and flexibility when compared to traditional generator-based approaches. We also describe the transformation of grammar flow analysis problems into an object-oriented framework using the Visitor pattern, as well as the implementation of a top-down LL(1) parser. As such, this work represents the synthesis of three presently disparate fields in parser design and implementation: combinator parsing, fixpoint-based grammar flow analysis, and object-oriented design.

Conflict analysis as a means of enforcing static separation of duty requirements in workflow environments

Perelson, S.Botha, R.A.

Abstract: The increasing reliance on information technology to support business processes has emphasised the need for information security mechanisms. This, however, has resulted in an ever-increasing workload in terms of security administration. Policy-based approaches have been proposed, promising to lighten the workload of security administrators. Separation of duty is one of the principles cited as a requirement when setting up these policy-based mechanisms. Different types of separation of duty policies exist. They can be categorised into policies that can be enforced at administration time, viz. static separation of duty requirements and policies that can be enforced only at execution time, viz. dynamic separation of duty requirements. This paper deals with specifying static separation of duty requirements in role-based workflow environments. It proposes a mathematical model based on the concept of “Sconflicting entities”T to express static separation of duty requirements. It provides a detailed explanation of the integrity checking that must take place at administration time to ensure that specified separation of duty requirements are honoured.

Defining and specifying graphs using formal model-based techniques

Kotze, P.

Abstract: Graph theory is an established field of study. The concepts of graphs and transition networks are well-known in computing. Mathematical expressions of some kind are almost always used to define graphs. Although these definitions are generally considered to be exact, one runs into difficulty when attempting to specify various graph definitions using a model-based specification notation such as Z, or when implementation structures are considered. In order to successfully do so one has to change the general mathematical definitions of graphs. This paper provides a set of such alternative definitions based on the use of bag structures.

Digital imaging challenges for artificial human vision

Boyle, J.Maeder, A.Boles, W.W.

Abstract: Current research to develop artificial human vision relies on the inducing of digital image-like effects in the brain. This paper reviews related projects underway internationally, and then discusses aspects of image processing which could provide enhanced visual information to users of such systems, where image quality is severely restricted. Some more advanced options involving computed pseudo-scenes providing virtual understanding of natural visual scenes are proposed.

Fast tracking students from disadvantaged backgrounds into main stream Computer Science

Blignaut, R.J.Venter, I.M.Cranfield, D.J.

Abstract: A computer-based training (CBT) system was used to teach Computer Literacy to full-time students at the University of the Western Cape. This approach was successful in creating computer literate students as well as creating an opportunity for students from educationally disadvantaged schools to enter the Computer Science course. The students experienced this new approach to learning positively. This has laid the foundation to export the computer-based education model to communities outside the university. Lifelong learning opportunities will thus be created.

Handling diversity in group Work in the information systems class

Thomas, T.De Villiers, C.

Abstract: Tertiary institutions in South Africa are faced with dealing with diversity in all its forms in our classrooms. Information Systems, Information Technology and Computer Science students need to learn to work with people who are different from themselves in order to learn to work effectively in the work environment to which they will go. Teaching students in a multicultural classroom to be able to practise their profession in multicultural settings is crucial. This paper looks at the problems that occur if we ignore diversity, some techniques for dealing with diversity, especially when using group work and then presents results of a series of four case studies where some of these techniques were applied.

Holistic programming environments

Marsden, G.Thimbleby, H.

Abstract: As a result of the popularity of graphical user interfaces, it is now almost impossible to buy a programming language compiler – instead, one purchases a development environment. Of course, we can scoff at the distinction and say that a development environment is nothing more than a programming language with visual (as opposed to syntactic) sugar. We believe, however, that this view must change if safer and more responsible programming languages are to be created for the next generation of programmer. Within this paper, we will argue that a more theoretical approach should be taken to the development of programming environments and suggest ways in which this may be achieved.

Organising educational resources on an Intranet

Min, Y.S.Thomas, T.

Abstract: Technology influences education in a number of ways. One of these influences is in the classroom, creating what can be called a computer supported learning environment. A computer supported learning environment has many tools available for the enhancing of the learning experience. Some of these tools are educational resources tempered with technology. These resources include interactive tutorials, animations and simulations. It is important for an educator to store these resources in such a way as to make them accessible to both students and fellow educators. In storing the resources, there are several factors that have to be remembered. These factors include the reusability of materials, the linking of materials to courses and the navigation through the course sites.

Presence as a means for understanding user behaviour in virtual environments  

Blake, E.Casanueva, J.Nunez, D.

Abstract: Presence has become a key concept in characterizing and evaluating Virtual Environments. Our contribution is to show that current measures of Presence, as a metric of users' experience of Virtual Environments, are highly problematic: results from the literature cannot be repeated and it lacks a theoretical basis. We synthesize results from three experiments we conducted and in conclusion point the way to alternative approaches to the problem of characterizing Collaborative Virtual Environments.

Virtual multicasting as an example of information mass transit

Machanick, P.Andrew, B.

Abstract: Virtual Multicasting (VMC) is a specific instance of a more general idea, Information Mass Transit (IMT). IMT aims to reduce the waste of bandwidth resulting from individual streams of data, while improving user-level latency. By analogy with mass transit where shared transport reduces the load on infrastructure, IMT aims to use networks and other infrastructure more efficiently. VMC combines some of the benefits of caching (transparency, dynamic adaptation to workload) and multicast (reducing duplicated traffic).