RA5a: School of Computer Science
The University has invested substantially in the School of Computer Science resulting in a 70% increase in Category A staff since 2001. Our strategy has been to recruit senior staff with outstanding research records to broaden and develop research within the School, together with an uncompromising approach to the recruitment of new junior staff with high potential to develop strong internationally competitive research activities. Large university and SRIF investments have provided a new building and infrastructure facilities for the School. The uplift in research output in quantity and quality is clear in terms of increases in research income (270% over the 2001 submission), research students and engagement in collaborative projects with other major research centres in the UK and abroad. Key changes since 2001 include:
1. In 2001, we returned our full complement of 12 academic staff; in this submission we are returning 21 category A staff.
2. A purpose-designed building (opened in 2004) provides state-of-the-art research facilities. All staff are co-located, strengthening the identity of the School and providing an excellent environment for communication and cooperation.
3. The number of research students (full-time + part-time) has increased from 19 in 2001 to 38 in 2007-2008.
4. Administrative, secretarial and technical staff have been appointed to research support roles.
5. In 2001-02, external research grants totalling £839,611 were awarded; in 2006-07, the corresponding figure was over £2.3 million.
The key phase of our expansion was in 2006 with the appointment of 3 new chairs and associated lectureships. Bhatti (from UCL), Bod (from Amsterdam), and Sommerville (from Lancaster) took up professorships with associated lecturers (Henderson, Nederhof, Bowles) joining us in 2006 and 2007. New grants and students associated with these appointments are now coming on-stream.
A cornerstone of our future strategy is the development of research within the Scottish Informatics and Computer Science Alliance (SICSA), an initiative by the Scottish Funding Council to encourage and support inter-university collaboration to create large, world-class research groups. We have had the leading role in formulating this initiative which is planned to start in early 2008. The university has made a firm commitment of a 7-year investment in computer science of £5.137m to supplement the requested Scottish Funding Council funding of £2.517m.
2 Research structure
Reflecting our recent expansion, we have reorganised our research under three broad themes:
A. Artificial intelligence and symbolic computation;
B. Networking and distributed systems;
C. Systems engineering.
We summarise the research in each theme, highlighting the contribution of externally funded projects and we discuss external impact, collaboration and inter-disciplinary links. Funding amounts quoted below reflect the award made to St Andrews, not the total award in collaborative projects.
As research impact is rarely traceable to individual pieces of work, our discussion focuses on how the body of work by St Andrews researchers has contributed to the development of its field and, where appropriate, how it has been taken up by industry.
2.1 Artificial intelligence and symbolic computation
Prof. S. Linton, Prof. R. Bod, Dr R. Dyckhoff, Dr I. Gent, Dr T. Kelsey, Dr I. Miguel, Dr M-J Nederhof, Dr B. Tiddeman, Dr M. Weir.
Research areas: computational algebra, constraint programming, computational logic, natural language processing, neural networks and image processing.
Linton, Kelsey, Gent and Miguel make up one of the strongest groups in the world working in the areas of computational algebra, symmetry and search and constraint programming.
Linton has an international reputation in computational abstract algebra, including the development of new algorithms and the application of computational algebra to problems in mathematics, physics and computer science. He is Director of the Centre for Interdisciplinary Research in Computational Algebra (CIRCA). CIRCA is funded by the EPSRC’s Multidisciplinary Critical Mass in Computational Algebra and Applications project (Linton, Gent, Miguel, £1,098,897) which focuses on group theory and which has established St Andrews as the leading world centre in using computers to work with symmetry. This project extends the work of the Symmetry and Search Network (EPSRC, Linton, £62,496), which investigated the relationships between computational search and symmetry.
Kelsey (early career) works closely with Linton in the area of computational algebra and, independently, is establishing an international reputation for his work with biomedical scientists in applying modelling techniques in cancer research. His extensive experience and skills bridging maths and computer science means that his appointment significantly strengthens our computational algebra group. He has been appointed to a lectureship but because of the importance of CIRCA, it has been agreed that he will continue as a senior research fellow until the end of the EPSRC CIRCA contract.
CIRCA has been involved, since 1997, in the development of the GAP system – an open source system for computational discrete algebra. This is now partially supported by the SCIEnce project (EC-FP6, Linton, Hammond, €658,260) that aims to create a symbolic computation infrastructure for Europe. Linton is a co-investigator, with colleagues in Mathematics and Statistics, on EPSRC-funded projects that rely on GAP (Semigroups and Monoids in GAP, £144,327; Applications of Automata and Languages in the Theory of Pattern Classes of Permutations, £157,723).
Gent and Miguel (early career), work closely with Linton and Kelsey, and are amongst the best known European researchers in constraint programming. Miguel joined us from York after completing his PhD in constraint programming, which was awarded a BCS Distinguished Dissertation award.
Their principal areas of work are in symmetry breaking, quantified constraints, automated constraint modelling and the design and implementation of efficient constraint solvers. Gent, Miguel and Linton explored symmetry in EPSRC-funded projects (Constraint Programming, Search and Symmetry, £146,610; Symmetry and Inference, £163,911) where they developed a new constraint modeling language and applied this to combinatorial real-world problems involving symmetry. The Refinement-driven Transformation of Effective Automated Constraint Modelling project, (EPSRC, £73,331) is extending this to explore techniques for the automatic generation of constraint models from an abstract specification. Miguel’s 5-year RAEng fellowship (£250,000), which focuses on the development of automated modeling, allows him to work full-time on this research. In a recently funded EPSRC project (Watched Literals and Learning for Constraint Programming, £382,018) Gent and Miguel will investigate the properties of ‘watched literals’ to explore how these can be used to increase propagation efficiency and to develop constraint programs that learn from their mistakes.
Bod and Nederhof form the core of a recently established natural language group where Bod has successfully adopted a statistical data-oriented perspective to natural language and music processing. Bod has recently received a prestigious ‘Vernieuwingsimpuls vici’ award (€845,000) for senior researchers from the Dutch Research Council, which will allow him to develop this work. Nederhof joined us from the Max Planck Institute for Psycholinguistics in 2006. He works with Bod in the area of formal models of natural languages as well as programming languages, with interests in generative power and complexity.
Dyckhoff has an international reputation for his work in computational logic, focusing on intuitionistic sequent calculi, including extensions of the depth-bounded calculus G4ip and the permutation-free sequent calculi that underlie proof search. He collaborates with Hammond (systems engineering) on the Embounded project (EC-FP6, €361,056) developing methods of reasoning about resource-boundedness of computations in embedded systems. His work in the Lexicalised Proof-Theoretic Semantics for Natural Language project (EPSRC, £59,304) follows a constructivist programme of semantics. Its goal is to develop techniques for natural language analysis that complement Bod’s statistical approaches to NL processing.
Tiddeman’s inter-disciplinary research (with Psychology) focuses on the synthesis and analysis of facial images, with applications in psychology and medical research. His EPSRC First Grant (Prototyping and Transformation of Facial Appearance and Dynamics in Video using 3D Deformable Models, £123,105) is developing a system capable of performing realistic alterations to video clips of moving faces, altering parameters such as the emotional state of the subject. Also in conjunction with the School of Psychology, he has industrial support from Unilever (£188,248) for work on the quantification of ageing and skin health from image analysis.
Weir’s research focuses on how neural networks, robots, and emotional agents produce goal-directed behaviour.
CIRCA is an international centre of excellence in computational algebra – its work on symmetry and group theory has influenced researchers around the world. GAP has a worldwide user base (around 1,200 sites) and, between 2000 and 2006, it was referenced in more than 800 refereed papers.
St Andrews is the world-leading site for research into the role symmetry plays in constraints. Our methods are incorporated into products, such as the constraint logic programming system ECLiPSE, developed and used by Cisco and Crosscore Optimisation. Our new, fast constraint solver, Minion, is used at the University of Toronto.
Dyckhoff’s research on the depth-bounded intuitionistic calculus has been heavily cited and is influential in the automated reasoning community with implementations inside major proof assistants like Coq, MetaPRL, Minlog and Isabelle.
The Data-Oriented Parsing model, developed by Bod, is one of the leading paradigms in statistical NLP, with applications in machine translation, music processing and unsupervised structure induction.
Tiddeman holds an EPSRC ‘Partnerships for Public Awareness’ award to demonstrate his work on facial transformations (Face of the future/explaining computer vision and graphics, £78,579). His interactive exhibit allows users to experiment with face transformation and explains the underlying computer vision and graphics technology. This has been demonstrated at 4 science festivals (York, Paris, Edinburgh, London) and installed as permanent exhibits in science museums in Dundee, Glasgow, Edinburgh and Manchester.
Academic and industrial collaboration
We are jointly responsible for managing the development of GAP with research groups at Braunschweig, Aachen and Fort Collins and collaborate closely in evolving this system. Linton collaborates with UWA on algebraic algorithms, Otago on combinatorics and Maplesoft. Dyckhoff collaborates with Helsinki and Paris 7, in the areas of computational logic and programming language semantics.
Gent and Miguel collaborate with researchers at Cork, York, Oxford, Glasgow, Uppsala and Imperial, on topics in constraint programming. Microsoft Research have provided funding for a studentship and we collaborate closely with Codeplay Ltd., who develop portable, high-performance compilers.
Bod works closely with the University of Amsterdam on natural language processing and Nederhof has established collaborations with the University of Bochum, on formal language theory, and the University of Padova on natural language parsing.
Inter-disciplinary collaboration: Our computational algebra research is inter-disciplinary involving colleagues in Mathematics and Physics. The development of GAP has involved mathematicians from various institutions around the world. Kelsey cooperates closely with medical scientists in Edinburgh University and the Royal Hospital for Sick Children, Edinburgh. Tiddeman works on facial dynamics and modelling with Unilever and with psychologists in St Andrews, Aberdeen, Stirling and Durham, with forensic anthropologists at Adelaide and Dundee and with orthodontists at Newcastle.
2.2 Networking and distributed systems
Prof. S. Bhatti, Prof. A. Dearle, Dr C. Allison, Dr T. Henderson, Dr G. Kirby, Dr A. Ruddle, Dr S. Norcross (RCUK academic fellow).
Research areas:Distributed and autonomic systems, network design and architecture, wireless and sensor networks and distributed learning environments.
Dearle works in distributed systems research including deployment, programming languages, middleware, file systems, and operating systems. Kirby’s interests are in distributed storage and the design and implementation of middleware systems. They work closely with Morrison and Balasubramaniam so that there is, in practice, a seamless interface between their work and work in systems engineering research.
The themes of operating systems, component deployment, and programming languages are combined in the DIAS-MC project (EPSRC, £249,128), reported in the systems engineering section. The Cingal project (EPSRC, £167,111) addressed the problem of creating distributed protection domains into which code and data could be injected. The Global Smart Spaces project (EC-FP5, €240,700) focused on devising software infrastructures to facilitate low-level interactions, driven by the explicit and the implicit high-level context of the user.
Work in autonomics includes the development of secure, location-independent storage architectures (EPSRC, £249,907) where we are investigating autonomic management in a peer-to-peer-based distributed file store to provide ubiquitous access to data with high resilience guarantees. This area of work is also supported by Norcross’s academic fellowship (RCUK, £125,000). The ASA infrastructure utilises novel middleware, developed in the Reflective Application Framework for Distributed Architectures project (EPSRC, £217,560) that allows instances of classes, written without regard to distribution, to be exposed as web services; programmers can control parameter passing semantics, object placement, and object migration.
Bhatti is a leading researcher in networked systems architecture and protocols, with particular interests in the control and management planes and their effect on systems’ behaviour and operation, and in network security. In IPv4 and IPv6 Performance and QoS - 46PaQ (EPSRC, £258,212) he carried out performance measurements of congestion control mechanisms, and new transport protocols operating under both IPv4 and IPv6. In the recently funded D-SCENT project (EPSRC, £168,251) he will investigate whether deception can be identified and proved from ‘scent trails’ – coherent accounts of suspects’ activities over time, compiled by tracking their movements, communications and behaviours.
Henderson (early career), who joined us from Dartmouth in 2006, cooperates with Bhatti in the area of wireless and mobile computing. His specific interests are in effective and efficient monitoring of mobile networks for improving network security and privacy. He was joint PI on an NSF project to set up an extensive archive of wireless data, CRAWDAD, at Dartmouth and is continuing and extending this work at St Andrews.
Allison and Ruddle’s are amongst the leading European research groups in distributed learning support. Their research interests are in the design, implementation and deployment of distributed learning environments and associated issues of network measurement and quality of service.
In the LeGE-WG: Learning Grid of Excellence Working Group (EC-FP5, €16,000) project, Allison investigated the potential of a grid-service based approach to distributed learning environments. This led to the ELeGI project (EC-FP6, European Learning Grid Infrastructure, €450,000) to develop such a grid. The Finesse project with the Dept of Accounting, Dundee (Stewart Ivory foundation, £42,000) developed a learning environment, widely used in finance teaching, which we use as a testbed for studying QoS in interactive, service-based applications.
The St-Andrews-Durham INSIDE project (An Institutionally Secure Integrated Data Environment, JISC, £180,000) identified the shortcomings of commercial products in creating a managed learning environment for academics; following this, the Scottish Funding Council supported the development of a repository for services as well as self-contained learning objects and static web sites (£53,000). Security is a critical issue in learning environments and we have developed a perception-based means of video identification that is robust against most transcodings of the types found in the copies circulating on P2P networks (Video Fingerprinting for Digital Assets Management, Scottish Enterprise Proof of Concept programme, £183,000). DSTL have funded a PhD studentship (£91,000) to investigate covert channels in public communications networks.
Dearle and Kirby work closely with Morrison as described below under systems engineering and have had a long history of work in persistence and reflection. Dearle’s long-term collaboration with Enigmatic has resulted in 2 patents:
1. Self-managed mediated information flow. US Patent US20040903156; 2004-07-30.
2. Mediated information flow. US Patent US2005060380; 2005-03-17.
The key innovation is the development of algorithms for managing a load balanced P2P network where a live system can autonomically change its deployment and topology whilst maintaining causal consistency with respect to externally observed behaviour.
Bhatti's work in high-speed networking included the first demonstration of software on a public network that handles dynamic QoS control at 2.5Gb/s. This software is a candidate for use on SuperJANET. His work with Henderson on user behaviour in networked games showed that assumptions about individual and group behaviour extrapolated from multimedia systems were invalid; many factors, including externalities beyond user control, affect user behaviour. Bhatti’s advisory work for OFCOM and ESA considering the future of wireless wide area networking for the UK using satellite networks (ESA) and terrestrial radio networks (OFCOM) has been used as input to national and EU policy.
Henderson is a co-investigator on the NSF-funded CRAWDAD wireless data archive, which has created the largest source of wireless network data, with 28 different data sets and 12 tools for processing data. It has 719 users from 414 different institutions and over 80 research papers have used CRAWDAD wireless data sets.
Allison and Ruddle’s work on distributed learning environments was instrumental in redefining the notion of distributed learning environments as service-based systems and they were early researchers in considering the importance of QoS and not simply user functionality. Two information systems, DIF and MMS, designed to support student learning and teaching have emerged from this research. These systems have been rolled out across the University as academic management information systems. They are currently being commercialized, in a University-funded project, with a view to launching them as products for the higher education sector in 2009.
Academic and industrial collaboration
Dearle and Kirby have a long-standing collaboration with Strathclyde, with industrial support from the Enigmatic Corp. Kirby collaborates with Abertay on software configuration tuning. Dearle was a consultant to Reuters Research and to Enigmatic from 2002-2005. He collaborated with Strathclyde University, Université Joseph Fourier, Trinity College Dublin and Starlab in the Gloss project.
Bhatti collaborates with: UCL, Edinburgh (Physics); Cambridge, Glasgow, and Loughborough. Industrial collaboration includes: UKERNA, Extreme Networks, Sun Microsystems and Microsoft Research. Topic areas are high-speed networking, network monitoring, network protocols & architecture, and network management. Consulting work includes collaboration with IABG Gmbh (Fritsch), ESA (multi-service satellite systems), PlexTek Ltd and OFCOM (wireless and mobile systems).
Interdisciplinary collaboration: Allison and Ruddle collaborate with Dundee (Accountancy) to maintain a realistic stock-trading environment for teaching and learning. They work with the OU Knowledge Media Institute and the LIRMM, Montpellier on service-based learning, ambient intelligence and enhanced presence, with Athens Information Technology on wireless network virtual laboratories and IBM Research, Zurich on Pervasive Computing.
2.3 Systems engineering
Prof. R. Morrison, Prof. I. Sommerville, Dr D. Balasubramaniam, Dr J. Bowles, Dr I. Duncan, Dr K. Hammond.
Research areas:software architectures and processes, dependable socio-technical systems and functional programming.
Our research in software architecture and processes integrates systems engineering and distributed systems research and involves Morrison, Balasubramaniam (early career), Dearle (NDS) and Kirby (NDS). Morrison has an international reputation for his work in persistence and programming language design. Balasubramaniam (early career) joined us as a lecturer after extensive experience as a researcher at St Andrews. Her appointment broadened our expertise in architecture and reinforced the strand of work linking architecture and software engineering processes.
Our principal interests are systems architectures for sensor nets, autonomic storage systems and the dynamic co-evolution of system structure and system behaviour. The work builds on a long history of programming language and persistent programming research where St Andrews has made major contributions. We continue to explore new approaches to software architecture and evolution through novel programming languages. For example, in the Archware project (EC-FP5, €370,809), we developedinnovative, architecture-centric languages, frameworks and tools for engineering evolvable software systems.
Our architecture and process work now focuses on the design, implementation and adaptation of sensor networks through multi-dimensional co-design and on autonomic systems research, discussed under the Networking and Distributed Systems theme. In the DIAS-MC project (EPSRC, £249,128), Morrison, Balasubramaniam, Dearle and Kirby are investigating how architecture and process co-design can result in sensor net systems that are optimal with respect to global cost functions specified by system designers. In the ASA project (EPSRC, £249,907), we are applying our ideas of architectural evolution to support the creation of a global persistent store.
Research in dependable socio-technical systems focuses on understanding and enhancing dependability by applying knowledge of a system’s social and organisational environment and in practical and theoretical approaches to improving the dependability of component-based systems. Staff involved are Sommerville, Duncan, and Bowles.
Sommerville is internationally known for his interdisciplinary research in how social, organisational and human factors influence the specification and design of systems. Over the last 15 years or so, he has developed a range of approaches designed to help software designers understand how social and organisational factors influence software dependability and use and has developed techniques to incorporate such knowledge in software engineering processes. In the INDEED project (EPSRC, £359,000), he is investigating how responsibility can be a useful constructive abstraction in designing complex distributed socio-technical systems and analysing these systems for potential vulnerabilities. The application domain being studied here is contingency planning. With support from an EPSRC Platform grant (EPSRC, £350,000), he is investigating dependability issues in e-elections and how responsibility modeling may be used to analyse these systems. This complements work on e voting by Duncan who has been concerned with developing usable and trustable systems for privacy and security related to e-voting systems (ESRC, £36,000).
In the recently-funded Large-Scale Complex IT Systems Collaboration (LSCITS) with Bristol, Oxford, York and Leeds (EPSRC, £1.15 million), Sommerville will extend previous work on integrating social analysis with systems engineering processes and explore new avenues of research examining socio-technical barriers to the use of innovative software systems, how people anticipate and cope with failure, and how systems can be designed to facilitate such coping behaviour.
Bowles (early career), who recently joined us from Birmingham, has research interests in formal approaches to systems dependability, complementing Sommerville’s socio-technical approach. She works in the general area of formal methods and software engineering, with a particular interest in the specification and design of component-based distributed dependable systems, and how formal approaches can be used to analyse and evaluate functional and/or non-functional properties of such systems.
Hammond is a leading researcher in functional programming and has made important research contributions in functional programming languages and environments (Haskell, Hume). His work now focuses on exploring higher-order parallel programming based on functional languages and in applying functional programming in domains with performance, time and space constraints. Projects include Generative Programming for Embedded Systems, (EPSRC, £145,065) and Automatic Prediction of Resource Bounds for Embedded Systems - Embounded, (EC-FP6, €361,056). An industrially-funded project with Heriot-Watt (BAe Systems, MoD, £187,740) is investigating the use of Hume in constructing real-time sensor analysis algorithms for autonomous vehicle guidance.
Hammond cooperates with the computational algebra group and is an investigator in the SCIEnce project (EC-FP6, €658,260) to create a symbolic computation infrastructure for Europe. This involves developing research prototypes supporting security, scheduling, and resource broking for computational grids and investigating the potential for exploiting functional programming in symbolic computation systems.
The long-term work in persistence (Morrison, Dearle, Kirby), where St Andrews is known as a world-leader, led to the development of mechanisms to implement persistence and reflection in a strongly-typed environment. These influenced persistence mechanisms, such as Hibernate and the CORBA persistence API. Our work on hyper-programming has influenced the design of BlueJ - a teaching environment for Java. Polymorphic addressing mechanisms, developed for Napier88, have recently been adopted in Microsoft's .Net framework. The need for access to compilers at run-time in a strongly typed environment, pioneered in PS-algol, has been included in Java V6. Strongly-typed structural reflection mechanisms, first provided by PS-algol, are now commonplace and are part of the .Net and Java frameworks.
Sommerville was a pioneer in involving social scientists in ethnographic studies of work to inform systems specification and design. He is continuing his work on socio-technical influences on system dependability (developed in the DIRC IRC), which established his group at Lancaster as one of the foremost world sites for socio-technical systems research. His research has contributed to recognition of the significance of socio-technical issues in dependable systems design (recognised, for example, in the funding of the DIRC IRC) and has influenced researchers working in requirements engineering and CSCW. There is increasing industrial interest in this area – companies such as Microsoft and Xerox have active groups in Europe, which include Sommerville’s former students or research assistants. Other companies, such as Rolls-Royce, are becoming interested in this area and have highlighted this as one of the key reasons for their involvement in the LSCITS project.
Hammond has had a long involvement with the development of Haskell, a widely used functional programming language. He has carried this work forward into the design and implementation of Hume, a strongly typed, mostly-functional language with an integrated tool set for developing, proving and assessing concurrent, safety-critical systems. St Andrews is one of a small number of sites investigating the application of functional programming in embedded systems development.
Academic and industrial collaboration
In software architecture, Morrison, Balasubramaniam and Kirby collaborate with Manchester, Bretagne-Sud and Adelaide with involvement from Massachusetts and Intel. With Dearle and Kirby, they collaborate with Manchester, Glasgow, Strathclyde and Kent on sensor nets research. Industrial partners include Xilink, Orange and Severn Trent Water. Morrison led the EPSRC Distributed Information Management network, with 13 UK partners.
Sommerville continues his collaboration from the Dependability IRC with Lancaster, York, City and Edinburgh, working in the area of socio-technical influences on system dependability. This involves Qinetiq, VOCA, NATS and British Nuclear as industrial partners. His work in using responsibility modelling to support contingency planning involves the Scottish Environmental Protection Agency. He is working in the LSCITS collaboration with Bristol, Oxford, York and Leeds. Industrial partners include Rolls-Royce, IBM and DSTL. Bowles works with colleagues in Birmingham, Surrey, Edinburgh and Versailles on theoretical models and foundations for component-based development taking into account aspects such as composition, concurrency, mobility, interaction and emergent behaviour.
Hammond’s involvement with the design and implementation of Haskell and, more recently, exploring the application of functional programming in embedded systems has led to collaboration with a wide range of institutions. Notable amongst these is Heriot-Watt, LMU Munich, Imperial College and Rice University. Industrial partners include BAe Systems, RECORE Ltd and Maplesoft.
Interdisciplinary collaboration: Sommerville has an established reputation in inter-disciplinary research. He has worked for almost 20 years with sociologists and management scientists in studying the influence of social and organisational issues on software engineering. His current collaborations in this area involve colleagues from the social sciences at Lancaster University and Edinburgh University and medical sciences at Leeds.
3 The research environment
3.1 Research students
We have seen a significant growth in research students between 2001 and 2008. In November 2001, we had 15 full-time and 4 part-time registered PhD students; by November 2007, this had grown to 34 full-time and 4 part-time PhD students.
We believe that the ideal ratio of research students to staff is about 2:1 as this gives us time to provide high-quality research training for our PhD students. Recently appointed staff have only just started establishing their PhD student groups so our immediate priority is to grow research student numbers to 40–45 to reflect the increase in academic staff.
The School invests in scholarships for first-class PhD students. In 2005-06, School funding covered studentships for 4 PhD students (EPSRC funding for a further 4 students). We have committed 4 internally-funded studentships as part of the India-UK partnership where fees will be remitted for 4 PhD students from India. The SICSA initiative will be an important source of funding (supporting 4 additional PhD students per year) and a means of attracting new PhD students.
3.2 Research income
Our expansion has led to a corresponding growth in research income and in the scale of externally funded projects. We anticipate that this upward trend will continue as EPSRC and FP7 proposals by the new staff appointed in 2006 and 2007 are developed and funded. A full list of awards is available from our website but, as an illustration, the value of awards announced in 2001-2002 was £839,611; in 2006-2007 the total amount announced was £2, 375, 270. The EPSRC is our principal research funder with additional support from the European Commission, DSTL, industry and other sources.
Sustainability of research income depends on active involvement of a body of academic staff in preparing research proposals, support for proposal development and a track record of success. Sixty percent (13 out of 21) of our Category A staff are listed as a Principal Investigator on one or more funded projects. We have established mechanisms for supporting research and the increase in income demonstrates our track record of success. Early career staff are encouraged and supported in developing research proposals through, for example. EPSRC’s First Grant scheme.
Increasing research income will naturally lead to an increase in the number of post-doctoral RAs in the School. Our intention is to increase from 17.5 to around 24 PDRAs by 2010 giving a 1:1 ratio of academic staff to PDRAs.
3.3 Research infrastructure
We have recently occupied a custom-built building incorporating state of the art laboratories and networking facilities. The building is equipped with a variety of formal and informal meeting spaces and has been designed as an experimental platform for pervasive computing with location sensors distributed throughout the building.
Research in network performance measurement is supported with specialised network measuring equipment, 48Tbyes of storage and an upgraded network capability offering 1 Gb/s to all desktops with some 10Gb/s connections. We have three Beowulf clusters for computationally-intensive tasks in computer algebra, distributed system simulation and video fingerprinting. A recently-installed network of 400 sensor motes supports sensor net research.
3.4 Technical and administrative support
Our technical and administrative research support has been significantly increased since 2001.We have appointed a full-time administrator who helps academic staff prepare research proposals and manages the administration of research grants; a part-time research administrator to support CIRCA and a dedicated research secretary. A research computing officer has been appointed who is responsible for supporting CIRCA research software.
3.5 Staff development
We are pro-active in supporting the development of senior and junior staff in that we plan our work so that staff have periods free from teaching where they can work exclusively on research, provide active support for research proposal development and mentoring support for early career researchers. We note here that 8 out of 21 Category A staff are members of the EPSRC’s peer college (almost 40%, one of the highest ratios in the UK) and this provides us with an experience base that we can call on to mentor early career staff.
Points to note here are:
1. School policy is that all staff have a semester free of teaching once every 4 years. This complements the University policy that allows staff to apply for additional research leave (1 semester in 8); strong applications for research leave are rarely refused.
2. Early career researchers are allocated lighter teaching loads for the first 3 years of their appointment to allow them time to develop as independent researchers.
3. Academic staff with significant research management responsibilities (currently Linton and Sommerville) are assigned a reduced teaching load.
4. A percentage of research grant overheads are returned to PIs to be used, at their discretion, for research support.
5. Staff are encouraged to apply for fellowship opportunities and cover their teaching accordingly. Several staff (Hammond, Kirby, Gent and Miguel) have been awarded EPSRC, Royal Society of Edinburgh or Royal Academy of Engineering fellowships.
6. Early career researchers are mentored within existing groups. Mentoring includes reviewing of papers and research proposals. Junior academic staff are encouraged to attend University-organised training courses (e.g. on research funding and proposal preparation).
7. Travel support is guaranteed for academic and research staff and students who have papers accepted in high-quality conferences.
4 Research strategy
Our strategic plan is to consolidate around our current research themes with new and replacement appointments strengthening these themes. After consolidation of our first phase of expansion, we plan a further significant increase in academic staff:
· An RCUK academic fellow (Norcross) has been in post since 2005 and has been appointed to a lectureship in distributed systems, which he will take up 2009. Kelsey, currently a senior research fellow in CIRCA, has been appointed to a lectureship in computational algebra which he will take up when current EPSRC support for CIRCA ends in 2010.
· Professors in distributed systems and in systems engineering will be appointed, subject to SICSA funding, in 2009.
· SICSA-funded lecturers will be appointed in all research themes in 2010 and 2011.
Our intention is that new appointments will be funded through the Scottish Informatics and Computer Science Alliance (SICSA). This grouping brings together leading Scottish computer science researchers in a collaboration led by Edinburgh, St Andrews and Glasgow Universities. A proposal for partial funding for SICSA is currently being considered by the Scottish Funding Council’s research pooling initiative that encourages the Scottish universities to pool their resources to create larger, world-class, research teams. A goal of SICSA, which reflects one of our own strategic priorities, is to attract top-class researchers from outside Scotland to new academic posts.
We have been closely involved in the planning and development of the SICSA proposal where we identified priorities for inter-university research collaboration that align closely with our research themes:
1. Next-generation Internet - aligned with networking and distributed systems;
2. Multimodal interaction - aligned with natural language processing;
3. Modelling and abstraction - aligned with constraint programming and computational algebra;
4. Complex systems engineering - aligned with systems engineering.
In line with SICSA objectives, we plan to develop further, pan-Scottish, collaborative research proposals. These will supplement existing collaborations in socio-technical systems with Edinburgh and in networking and sensor nets with Glasgow.
Across Scotland, investment in SICSA will be more than £30 million over 7 years, with around £14 million in additional funding from the SFC. This will fund additional academic staff and the creation of the Scottish Graduate Academy (SGA). The SGA will fund research studentships, academic visitors, shared research supervision and postgraduate summer schools. It has been provisionally agreed that Prof. Sommerville will be released from teaching duties to become the first Director of the Scottish Graduate Academy (20% FTE).
Locally, the University has agreed to contribute £5 million+ to support the SICSA initiative from 2008-2014. This is in addition to SICSA funding for St Andrews in excess of £2.5 million.
An important strategic goal for us is to increase PhD student numbers. Recent appointments are contributing to this and we anticipate significant further expansion over the next 5 years. The SGA will provide 50% of the funding for up to 4 students per year for 5 years, so that we will double the number of PhD studentships that we provide from our own resources. Furthermore, this will help increase our international profile and hence attract further self-funded students.
5 Esteem indicators
A: Artificial intelligence and symbolic computation
Prof. S. Linton
Best paper award: IEEE Visualization 2003, Seattle.
Journal editor: Applicable Algebra in Engineering, Communications and Computing, Springer.
Member: EPSRC Strategic Advisory Team, 2006 –.
Dr T. Kelsey (early career)
Invited speaker: Annual Scientific Meeting of the Sindicato Médico del Uruguay, 2007.
Conference organiser: International symmetry conference, Edinburgh, 2007.
Dr R. Dyckhoff
Senior Visiting Fellow: Institute of Advanced Studies, Bologna, 2005.
Consultant editor: Studia Logica, 2004 –.
PC Member, Conference series: Automated Reasoning with Analytic Tableaux, 2001–.
Dr I. Gent
Associate editor: J. of Satisfiability, Boolean Modelling and Computation, 20042007, IOS Press.
Editorial Board Member: Constraint Programming Letters, 2006 –.
PC member: 19th Int. Joint Conf. on Artificial Intelligence, Edinburgh, 2005.
Dr I. Miguel (early career)
Distinguished Dissertation: PhD Thesis awarded BCS/CPHC Distinguished Dissertation Award, 2002
RAEng/EPSRC Research Fellow, 2004-09.
Co-chair: SARA'07 7th Int. Symp. on Abstraction, Reformulation and Approximation, Whistler, Canada, 2007.
PC member: 17th European Conf. on Artificial Intelligence, Trento, Italy, 2006.
Prof R. Bod
‘Vernieuwingsimpuls vici’ award (€845,000) by NWO (Netherlands), 2007.
Symposium organiser: Probability Theory in Linguistics; Washington, 2001, Atlanta, 2003.
Dr M-J Nederhof
PC member/Area Chair, Formal NLP: HLT/EMNLP 2005, Vancouver, 2005.
PC member/Area Chair, Parsing: COLING-ACL 2006, Sydney, 2006.
B: Networking and distributed systems
Prof. S. Bhatti
Director and PI, National Centre of Excellence in Networked Systems, 2002–2005.
Keynote speaker: Questnet Networking Conference, Coolum, Queensland, Australia, 2005.
Guest editor: J. of Grid Computing, special issue on High-Speed Networking, December 2003.
Prof. A. Dearle
Program co-chair: 3rd Int. Working Conf. on Component Deployment, Grenoble, 2005.
Keynote speaker: 3rd UK Ubinet Workshop, Bath, 2005.
Dr G. Kirby
Awarded Royal Society of Edinburgh Personal Research Fellowship, 2005–2006.
PC Member, Conference series: IEEE Consumer Communications and Networking Conf., 2003–.
Dr C. Allison
Keynote speaker: TICE 2006: Information and Communication Technologies in Higher Education and Industry, Toulouse, 2006.
Keynote speaker: IWI2006 – 2nd Workshop on Innovations in Web Infrastructure, Edinburgh, 2006.
Tutorial Chair and PC member for the 7th IEEE International Conference on Advanced Learning Technologies (ICALT 2007), Niigata, Japan.
Dr T. Henderson (early career)
PC member: 4th IEEE Conference on Sensor Mesh and Ad Hoc Communications and Networks, San Diego, 2007.
C: Systems engineering
Prof. R. Morrison
Fellow of the Royal Society of Edinburgh, 2001.
Co-chair: European Workshop on Software Architecture, 2004, 2005 (St Andrews, Pisa).
Keynote speaker: European Conference on Software Architecture, Madrid, 2007.
Prof. I. Sommerville
General Chair: IEEE COMPSAC 2002, Oxford, 2002.
Keynote speaker: 6th Int. Conf. on eXtreme Programming and Agile Processes in Software Engineering, Sheffield, 2005.
Keynote speaker: 21st IEEE Int. Conf. on Software Maintenance, Budapest, 2005
Keynote speaker: 1st IEEE Conf. on Digital Ecosystems, Cairns, Australia, 2007.
Dr D. Balasubramaniam (early career)
PC member, Workshop series: European Workshop on Software Architecture, 2004–2006.
Dr J. Bowles (early career)
Programme Chair: International Workshop on Formal Foundations of Component-Based Software Architectures and Embedded Software (FESCA). 2004, 2005 and 2007.
Awarded Ramon y Cajal Research Fellowship by the Spanish Ministry of Science and Technology 2004-2009 (not accepted because appointed to a permanent UK academic post).
Dr K. Hammond
Awarded Royal Society of Edinburgh Personal Research Fellowship, 2006–2007.
General Chair: 6th Symposium on Trends in Functional Programming (TFP 2005), Tallinn, Estonia, 2005.
Invited speaker: Central European Summer School on Functional Programming, Budapest, 2005.