FAULT ANALYSES IN INDUSTRY AND ACADEMIC RESEARCH

Faults in software cost the UK IT industry hundreds of millions of pounds each year; a common aim of software engineering research in academic terms is to address industry problems. The Fault Analyses in Industry and Academic Research Network (FIAR-NET) brings closer ties between industry and academic research in the area of software fault analysis. A range of pressing issues exists and it is to the mutual benefit of the two communities that they work more closely than they have in the past. This is the underpinning motivation and reason for proposing FIAR-NET.

Two key themes run through the network. Firstly, challenges of fault case-studies using industrial systems. While recent progress has been made in this area, more industrial involvement has the potential to enhance and add high value to these foundations. The theme will embrace the practicalities of predicting faults, fault data collection/assimilation (including ensuring that the data is reliable), fault data interpretation and faults in an 'agile' context, a model for developing software which is prevalent in UK IT industry. The second theme is an exploration of approaches to re-engineering code so that is in some way better and thereby mitigate and minimise faults; finally through better testing strategies (both pre- and post-release).

FIAR-NET will operate through regular workshops, events and a range of on-line resources and initiatives including a set of on-site "road-shows". An initial network of core academics and industrial partners including SMEs and large organisations will be established, maintained and then extended. A key aim is to stimulate innovative, leading research collaboration and dissemination of subsequent results. The proposed network team at Brunel builds on a track record of expertise in fault analyses and empirical software engineering, widely recognised as world-leading. This reputation is underpinned by a wide variety of past and present industrial and academic collaborations which FIAR-NET will build on in the three years of its duration and thereafter.

Both Counsell (PI) and Hierons (a core FIAR-NET member) have led successful research networks in the past and all investigators on the project have wide experience of industrial collaborations and bring that experience to the network. Finally, FIAR-NET falls within one of the EPSRC's strategic aims and portfolio plan to 'grow' research into fault-finding and fixing.

ADVANCEMENT OF KNOWLEDGE

Advancement of knowledge facilitated by FIAR-NET can be seen from four viewpoints. Firstly, the areas that FIAR-NET explores are open research areas. For example, the optimal set of characteristics of code and combinations of code features for predicting faults has yet to be established. Equally, the link between faults and refactoring and whether refactoring actually improves the quality of code has received very little empirical attention. Progress in each of these areas has been hampered by the lack of data from proprietary systems and so involvement of industry is vital to understanding and then solving some of these problems. The proposed network brings together an initial set of six industry partners, the BCS and four academic institutions, all with specific interests in these types of area. Secondly, the results can be adopted, adapted and pursued by industrial and academic collaborations through the project deliverables; specifically, through the workshop agendas and road-shows, both of which provide a forum for research to be further explored. The set of dissemination activities and potential for extending the network over the project duration also promote knowledge advancement. Thirdly, the research agenda will inform new knowledge about, a) collection and analysis of fault data, b) testing, the problems, differences and emphases of commercial test environments (Agile or otherwise) and c) how systems can be re-engineered to minimise faults. The project thus provides many opportunities for better allocation and prioritisation of resources by developers and project managers.

PEOPLE IMPACT

From a people perspective, the project promotes a set of skills through the investigation of research problems shared between industry and academic groups. These skills can be adopted by practitioners who wish to apply the same principles to their software and fault bases. A range of people will also benefit from the research. Software architects, managers and developers at the industrial partners will benefit directly through results and lessons generated from the network. The network will allow academic researchers in the areas that FIAR-NET covers to benefit from a shared understanding and the transfer of knowledge that academic and industrial collaboration fosters.

ECONOMIC IMPACT

From an economic viewpoint, faults in software will always occur and there will always be some cost associated with faults and fixing them when they do occur. However, there are significant potential cost savings to be gained by the UK IT industry if research can produce demonstrably successful methods of understanding faults and tackling issues surrounding faults. Firstly, resources can be pre-allocated and directed where they are most needed on a proactive rather than a reactive basis, the former of which often causes disruption to other projects. Secondly, explanations of the causes of faults can improve and streamline the processes through which software is developed. Finally, the ability to understand system behaviour in a more timely fashion ultimately frees up time for developers to work on improving/perfecting the quality of their code. The proposed research has huge potential for unlocking the opportunity that a more timely awareness and understanding of system pressures create.

SOCIETAL IMPACT

From a societal viewpoint, the research is in alignment with a key aim of any company of developing robust code and high quality products. Customers of the industrial partners will benefit in terms of the software that they 'deliver'. The value of the research also appeals to the wider societal community. In fact, the research affects anyone who uses software since improvement in the quality of code through minimisation of faults is a fundamental aim of any software developer.