South African Computer Journal - Volume 57, Issue 1, 2015
Volumes & issues
Volume 57, Issue 1, 2015
Source: South African Computer Journal 57, pp VII –VIII (2015)More Less
SACJ has had a layout designed for print up to now. For print, if you are trying to save paper count and therefore use a small font, it becomes difficult for the eye to follow a long line. That is why multiple columns were invented. Now that the journal is published online, there is no strong reason to maintain a layout designed for print with a small font and two columns. Two columns make for difficult reading on a small screen as you have to read to the bottom of a left-side column then scroll back up for the next column. Two-column layout also makes it hard to fit in decent-sized figures and tables, leading to a compromise of allowing them to run across both columns if necessary. Going to single-column layout removes all these problems.
Syllabification and parameter optimisation in Zulu to English machine translation : research articleSource: South African Computer Journal 57, pp 1 –23 (2015) http://dx.doi.org/10.18489/sacj.v0i57.323More Less
We present a series of experiments involving the machine translation of Zulu to English using a well-known statistical software system. Due to morphological complexity and relative scarcity of resources, the case of Zulu is challenging. Against a selection of baseline models, we show that a relatively naive approach of dividing Zulu words into syllables leads to a surprising improvement. We further improve on this model through manual configuration changes. Our best model significantly outperforms the baseline models (BLEU measure, at p < 0.001) even when they are optimised to a similar degree, only falling short of the well-known Morfessor morphological analyser that makes use of relatively sophisticated algorithms. These experiments suggest that even a simple optimisation procedure can improve the quality of this approach to a significant degree. This is promising particularly because it improves on a mostly language independent approach at least within the same language family. Our work also drives the point home that sub-lexical alignment for Zulu is crucial for improved translation quality.
Author Michael MitchleySource: South African Computer Journal 57, pp 24 –31 (2015) http://dx.doi.org/10.18489/sacj.v0i57.284More Less
Reinforcement learning is a machine learning framework whereby an agent learns to perform a task by maximising its total reward received for selecting actions in each state. The policy mapping states to actions that the agent learns is either represented explicitly, or implicitly through a value function. It is common in reinforcement learning to discretise a continuous state space using tile coding or binary features. We prove an upper bound on the performance of discretisation for direct policy representation or value function approximation.
Virtual learning system usage in higher education - a study at two South African institutions : research articleSource: South African Computer Journal 57, pp 32 –57 (2015) http://dx.doi.org/10.18489/sacj.v0i57.324More Less
In higher education institutions various VLSs have been formally adopted to support online teaching and learning. However, there has been little research on patterns of VLS use among educators. The purpose of the research was to provide a descriptive analysis of VLS feature usage, and associated challenges at two South African higher education institutions. A case study research strategy was adopted, combining qualitative and quantitative approaches to data collection and analysis. Survey findings revealed four clusters of VLS feature usage, namely, communication, management, content and pedagogic. Analysis showed that the 'content cluster' was used more than the other clusters. The average usage of the 'pedagogic cluster' for Durban University of Technology (DUT) was significantly greater than that of University of KwaZulu-Natal (UKZN), which tentatively indicates that staff development seems to be an important aspect of VLS usage. There was no significant difference in the usage of the 'communication' and 'management' clusters between the two institutions, DUT and UKZN. The study contributes to the body of system utilisation research by confirming an uneven pattern of VLS feature usage among educators, whilst providing fresh insights into the challenges associated with the usage of two different VLSs in two different universities.
Source: South African Computer Journal 57, pp 58 –89 (2015) http://dx.doi.org/10.18489/sacj.v0i57.294More Less
This paper studies the output of South African universities in terms of computing-related doctorates in order to determine trends in numbers of doctorates awarded and to identify strong doctoral study research areas. Data collected from a variety of sources relating to Computing doctorates conferred since the late 1970s was used to compare the situation in Computing with that of all doctorates. The number of Computing doctorates awarded has increased considerably over the period of study. Nearly three times as many doctorates were awarded in the period 2010-2014 as in 2000-2004. The universities producing the most Computing doctorates were either previously "traditional" universities or comprehensive universities formed by amalgamating a traditional research university with a technikon. Universities of technology have not yet produced many doctorates as they do not have a strong research tradition. The analysis of topic keywords using ACM Computing classifications is preliminary but shows that professional issues are dominant in Information Systems, models are often built in Computer Science and several topics, including computing in education, are evident in both IS and CS. The relevant data is in the public domain but access is difficult as record keeping was generally inconsistent and incomplete. In addition, electronic databases at universities are not easily searchable and access to HEMIS data is limited. The database built for this paper is more inclusive in terms of discipline-related data than others.
Synchronous gesture manipulation for collaboration and coordination of co-located business process modelling : research articleSource: South African Computer Journal 57, pp 90 –112 (2015) http://dx.doi.org/10.18489/sacj.v0i57.269More Less
The purpose of this study was to investigate approaches (techniques and technologies) for the coordination of collaborative tasks using synchronous gesture manipulation. Business Process Modelling (BPM) tasks are often performed in teams of modellers who need to collaborate with each other in order to coordinate and integrate their individual contributions into the various process models in a co-located environment. These collaborative BPM tasks were used as a case study in order to develop the artifact (the BPM-Touch approach)as a proof of concept. The BPM-Touch approach allows for the coordination and collaboration of BPM tasks in co-located modelling teams using synchronous gesture manipulation approaches. The Design Science Research (DSR) methodology was used and several cycles of developing and evaluating the artifact took place. This paper reports on the last cycle and set of evaluations. The proposed approach was implemented in a BPM software package in order to provide empirical validation. Usability evaluations of the software were undertaken with both students and BPM professionals as participants. The empirical results of the evaluations revealed that the participants found the approach to be effective and rated the usability and satisfaction of the collaboration and gesture manipulation aspects of the software positively.
Source: South African Computer Journal 57, pp 113 –117 (2015) http://dx.doi.org/10.18489/sacj.v0i57.347More Less
The use of graphics processing units (GPUs) in general-purpose computation (GPGPU) is a growing field. GPU instruction sets, while implementing a graphics pipeline, draw from a range of single instruction multiple data stream (SIMD) architectures characteristic of the heyday of supercomputers. Yet only one of these SIMD instruction sets has been of application on a wide enough range of problems to survive the era when the full range of supercomputer design variants was being explored : vector instructions.