Mutation Testing Repository [Visit]
Mutation Testing is a fault–based software testing technique that has been widely studied for over three decades. The literature on mutation testing has contributed a set of approaches, tools and empirical studies. This repository aims to provide a full coverage of the publications in the literature on Mutation Testing.
A utility tool for finding suitable configurations for clone detection tool
A data set which has been used for our App Store Mining and Analysis project
A mutation testing tool
Evo-Bentham (Joint work with Bill Langdon, Mark Harman and Yuanyuan Zhang, submitted to UCL Research Images Competition 2014)
Description: This picture shows seven portrayals of Jeremy Bentham, constructed using transparent polygons by Genetic Programming. Genetic Programming is an evolutionary approach that is applied to evolve computer programs, guided by the principles of Darwinian evolution. Here we apply this approach to evolve pictures from a set of arbitrary polygons (many side shapes) with different colour and opacity. Our approach tries to simultaneously favour both of two conflicting objectives when deciding on the desirability or each picture: 1) Does it capture the essence of Jeremy Bentham?, which we measured as the difference in colours at each point in the picture; and 2) Is it artistically abstractionist?, which we measured as the number of polygons needed to draw the picture (the fewer the more abstract). We seek a trade-off between minimising the difference in colours (improving the likeness) and minimising the number of polygons (increasing the abstraction). Our portrayals have been selected from a so-called `Pareto' front, which contains solutions that trade off these two objectives. Whilst the leftmost portrayal only uses a single polygon (so it is maximally abstract), the rightmost portrayal uses 60 polygons (it is less abstract but also it is a closer likeness to Jeremy Bentham). Which portrayal do you prefer?
Software, like trees, grows and evolves (submitted to UCL Research Images Competition 2013)
Description: Software, like trees, grows and evolves all the time. Software architectures are typically visualised as directed graphs. Such representations are precise, but it is difficult to conceptualise the entire software in this way. This project introduces a new method for visualising software architectures in a more natural and realistic tree form. Here in this image, the popular compression software GZIP is compared at two stages of its development: version 1.2 (in 1993) and version 1.6 (in 2013). The trunk of the tree symbolises the main program function and the entry point for its execution. Each branch of the tree represents the instance at which a subroutine in the program is referenced and utilised. The new diagram provides an elegant way to discover how GZIP has grown over the past 20 years.