Bertram Haskins

Evaluating plagiarism detection software for introductory programming assignments

Plagiarism is an issue that all educators have had to deal with. Large numbers of students and assignments have resulted in the development of automated systems to detect code similarities with the aim of identifying cases that may have been plagiarised. These systems are of great value to assessors, allowing them to process submissions automatically. However, these automated systems do present possible disadvantages and drawbacks. In this study we explore and analyse the differences between various systems as well as how their performance compares with manual checking. We consider the different methods students use when committing plagiarism. Then we examine more closely the systems that can aid plagiarism detection, ranging from their characteristics to how they work. In the process, we determine how these systems compare with our own system and their suitability for aiding the identification of submissions which may have been plagiarised in our introductory C++ course.

Claimed vs observed information disclosure on social networking sites

Research on internet users reports a gap between reported privacy concerns and observed privacy behavior. This gap has also been reported on social networking sites like Facebook and MySpace. This current study explores the extent of this gap by examining individual Facebook components. Unlike the previous studies that have explored reported privacy concerns versus observed behavior, this study explores claimed information disclosure versus observed information disclosure. A questionnaire was used to examine the reported information disclosure and observations on public Facebook profiles were conducted to examine the actual behavior. The study shows that there is a gap between the reported information disclosure and observed disclosure, like in the case of reported privacy concern versus observed behavior. However, this study shows that the users act more securely than they claim to act.

Factors that Influence the Choice of Privacy Settings on Facebook: Freshmen's View at a South African University

A contradiction between the reported privacy concerns and observed privacy behaviour of users of social networking sites has been reported. To understand this contradiction this paper examines factors that influence privacy behaviour of users on social networking sites. Participants in this study were freshmen studying IT recruited from Nelson Mandela Metropolitan University during the academic year of 2011. A questionnaire was used to study the reported privacy concerns of the participants, followed by observations on the public Facebook profiles of the participants to shed light on their actual behaviour. Interviews exposed factors that negatively influence privacy behaviour, while another questionnaire explored factors that positively influence privacy behaviour. This study shows that the contradiction between privacy concerns and observed privacy behaviour varies according to the type of information in question. Based on the findings, we developed a model that identifies factors that users consider when choosing privacy settings on Facebook.

Devising NREN design goals from commercial network design guidelines

To design a network as big as SANReN network design goals need to be determined. The network design process is an iterative one that follows a top-down approach. The network design lifecycle is a continuous process that seeks improvement of the network design so that it better addresses business needs and goals. This paper provides an overview of the network design goals model that provides a process for determining network design goals for an NREN. The model is instantiated in the SANReN environment. The network design goals may be slightly different to a commercial service provider but the process for determining network design goals for a commercial service provider can be adapted for SANReN.

A mixed-method approach to normalising Dr. Math microtext

Dr. Math is a South African math tutoring service, which allows high school learners to use a mobile phone-based chat application to contact volunteer tutors. The learners structure their queries in a form of microtext, which the tutors find difficult to interpret. This study presents an automated process as a means to normalise a Dr. Math microtext-based learner query, so that it more closely resembles English. The process consists of a variety of replacements, based on known short forms, pre-generated English-variants and phonetic similarity. Input queries are processed at a word-level, regardless of underlying grammar. The normalisation process is validated in two steps. The first step uses spell-checking to determine that the process improves the overall spelling accuracy, of the original microtext-based input queries, in both test and validation conditions. Krippendorffs α is used in the second validation step to calculate the level of agreement, with regards to the legibility of normalised text, between two human coders. The α value of 0.681 is deemed sufficient for exploratory research. The coders agree that the normalisation process yields an improvement in legibility.