You are in: Submissions > Select unit of assessment > UOA 23 Computer Science and Informatics > University of Kent > RA5a

UOA 23 - Computer Science and Informatics

University of Kent

RA5a: Research environment and esteem

Group Structure

The Computing Laboratory has six research groups: Applied and Interdisciplinary Informatics (AII), Cognitive Systems (CogSys), Computing Education (CompEd), Information Systems Security (ISS), Systems Architecture (Systems) and Theoretical Computer Science (TCS).

Note: superscript "e" signifies names of Early Career Researchers.

Achievements

This section describes achievements concerned with infrastructure and process; scientific successes are described by research group below.

In the period 2001-7, the Laboratory has increased its returnable annual research income from £226,520pa to £984,624pa, a factor of over 4; returnable research postgraduate numbers have increased from 23 to 39, a factor of nearly 2, stimulated by bursaries pump-primed by the Laboratory and the University, and we have graduated 56 research students. These improvements have markedly increased the department's research vitality.

We have maintained our strongest, most mature research groups, while, at the same time, encouraging fresh research activities, e.g. the IS Security group led by Chadwick. We have fostered interdisciplinary activities, particularly via the AII group, with close links to Biosciences and other departments through the new Centre for Biomedical Informatics (Deputy-director Johnson). We have also established a separate Cognitive Systems group, supporting our recent launch of the joint Centre for Cognitive Neuroscience and Cognitive Systems (CCNCS) with Psychology (directed by Bowman). These centres form natural foci for the development of new research programmes and activities. Award of two of the first RCUK Academic Fellowships to the Laboratory (Chue and Cowelle) is clear evidence of the success of our interdisciplinary activities.

We have also established a wide range of collaborations with other institutions; examples leading to significant funding include Thompson with Derrick at Sheffield (FP7 ProTest project), Welch with Timmis at York (EPSRC CoSMoS project), Rodgers with Howse at Brighton (EPSRC Reasoning with Diagrams project).

Strategy to 2013

We will maintain the existing groups, providing continuity of research activities, and recruit staff so that smaller groups have critical mass and sustainability. Some 23% of our submitted staff are early career researchers, reflecting our sustainable research base.

We have recently concentrated expansion on the ISS group, building links with the various aspects of security already covered within the Systems and TCS groups, by recruiting Hassene as Lecturer in security. This meshes with the University's strategic priority to build institution-wide research and enterprise activities in security.

We will continue to increase and broaden our research funding, with a Departmental target for awards of £3.5Mpa by the end of 2013 (45% above the current rate of £1.96Mpa). For research postgraduates, the aim is to stabilize total numbers at about 60, but with a greater proportion of external funding; the increased DTA funds resulting from our growth in grant income helps do this. We aim to foster further external collaboration by targetting FP7 calls; already, two substantial grants have been awarded (Chadwick and Thompson). We will consolidate collaboration within the University by strengthening our interdisciplinary research centres, and use this strength to promote participation in new national and international initiatives, e.g., using the capabilities of the CCNCS in national Neuroscience initiatives.

The University has scheduled major investment (provisionally £6.2M) in new custom-built accommodation for Computing in the 2008-2010 period, so the physical environment will soon be significantly improved.

Further detailed plans appear under each research group.

Collaboration and Dissemination

The Laboratory has a wide range of links with industry, other disciplines, Computer Science in other institutions and with the general public.

Our substantial industrial links include recognition of our Computing Education Research by award of the Sun Center of Excellence in Object-Oriented Education and as the first Sun Target University outside the US, with associated funding for Kölling and Utting. Funding from Microsoft has included a Rotor award for Jones's Memory Management Visualization; Robert Berry of IBM is a Visiting Professor, and both Jones and Linington have been granted IBM Faculty Awards. Work on refactoring by Thompson has lead to FP7 collaboration with Ericsson. There are strong links on security, by Chadwick with Nexor (defence and government security specialist) and by King with Portcullis Computer Security (security testing and management house). Other industrial partners include Agilent, BAe, BT, Citrix and NCR.

We encourage Laboratory staff to organize international research conferences at Kent. Examples are ITiCSE'01 (Fincher), the invited workshop on Engineering e-Business Systems 2002 (Kent, de Lemos and Linington), IFM2004 (Boiten), Second European Workshop on Model Driven Architecture 2004 (Akehurst, Linington), MM-net Summer School 2004 (Jones), 1st International Conference on Artificial Immune Systems 2002, Second EuroPKI Workshop 2005 (Chadwick), European Conference on Artificial Life: ECAL2005 (Freitas, Johnson), ACM ICER-2006 (Fincher) and UKKDD-2007 (Freitas).

Technology transfer is delivered by Thompson's recent KTP award for Refactoring in the Erlang language and in collaborations with Axon on remote healthcare systems. In outreach to the broader community, there is a strong programme to promote computer science in local schools through workshops based on the greenfoot and BlueJ environments developed by Kölling and others, and Chue, an Academic Fellow, has established a local "Cafe Scientifique", in which research is presented informally to open audiences in a local pub.

Service to the International Community

In addition to direct involvement in collaborative projects, the laboratory has a large involvement in peer review and conference organization. As well as journal and grant reviewing, five submitted staff have been EPSRC College members since 2001. Over the period, the Laboratory has provided 60 external PhD examiners (29 overseas), 25 journal editors, and 320 conference programme committee (PC) members, including 60 PC chairs.

Departmental evidence of esteem

1

Award of two of the first tranche of RCUK Academic Fellowships (Chue and Cowelle).

2

First Centre of Excellence to be sponsored by Sun Microsystems outside the USA (2006).

Individual research groups

Research in the Computing Laboratory is sustained, managed and developed through its six research groups which meet regularly, with events ranging from technical presentation by external or internal speakers to reading groups, planning sessions identifying new opportunities, or reviews of project proposals from members. All groups maintain a web presence (http://www.cs.kent.ac.uk/research/) recording work in progress and information on new initiatives or previous projects.

Some of our flagship groups, such as TCS, represent evolving activity over more than twenty years. Others represent newer initiatives. They offer an efficient working structure, encouraging the exchange of best practice, and a collaborative environment for testing and developing new ideas. The structure responds to shifts in interest or changes in staffing. A major review in 2004 introduced several new groups, triggered by the arrival of staff; minor adjustments continue to be made. Each staff member has a home group, but can be a member of others if appropriate; cross-group activities are encouraged.

Staff are listed by group, distinguishing primary affiliation (used for reporting statistics) from associate affiliations (indicating activity spanning groups).

Applied and Interdisciplinary Informatics

Primary staff: D.F.Chue, A.A.Freitas (head), C.G.Johnson, P.J.Rodgers, A.R.Runnalls

Associated staff: N.S.Ryan

Income £1,011,732; 17.5 research degrees awarded; 17.4 current students + 1 writing-up.

The AII group is involved in interdisciplinary research integrating computer science and areas such as biology, mathematics, astronomy, anthropology and healthcare. The objective is to develop new computational methods, supporting researchers in a wide variety of related areas. This involves two major research threads, natural computation and visualization.

Natural computation involves Freitas, Johnson and Chue. Freitas is developing new bio-inspired algorithms for challenging data mining and bioinformatics problems, such as hierarchical, multi-label classification and protein function prediction. He has developed new algorithms in ant colony optimization [Freitas:4], particle swarm optimization, genetic algorithms, genetic programming [Freitas:2, Freitas:3] and artificial immune systems [Freitas:1]. He has published two research monographs and more than 100 refereed papers. He has been the principal investigator in three EPSRC grants and one Interreg (Franco-British) grant involving bio-inspired algorithms for data mining and/or bioinformatics.

Johnson's research involves three main areas: first, developing computational and mathematical methods to support research in biology [Johnson:2-3], medicine and healthcare; secondly, taking ideas from nature and using them as metaphors in computer science, particularly for understanding and managing complexity [Johnson:1,4]; thirdly, developing computational methods for music and media technologies. He has been a co-investigator in three of the four grants held by Freitas, and in Bowman's EPSRC computational neuroscience grant.

Chue is an Academic Fellow, whose main research interests are bio-inspired computing and systems biology. He works on computational systems that explain the evolution of cell signaling networks with pre-specified properties [Chu:2-3]. He collaborated with biologists in modelling a gene expression pathway in E. Coli [Chu:1]. He has also worked on the underpinnings of complexity and artificial life [Chu:4].

Freitas, Johnson and Chue all have strong collaborations with Biosciences at Kent, and have all published refereed journal papers co-authored by biologists. They are involved in the interdisciplinary Centre for BioMedical Informatics; Johnson is the Deputy Director and Freitas serves on its Steering Committee.

Visualization is led by Rodgers, whose research interests focus on diagrams representing complex interactions between objects, such as Euler Diagrams [Rodgers:1,2] and Graphs. He develops practical layout applications (such as automated Metro map layout and WWW visualization), as well as graph rewriting programming languages and diagrammatic reasoning systems. He has applied visualization tools to other application areas [Rodgers:4]. Rodgers has been the principal investigator in three EPSRC grants relating to graph rewriting for graph drawing, diagrammatic reasoning for software modelling and visualizing Euler diagrams. He has been co-investigator in an EPSRC grant on multi-viewpoint visualization [Rodgers:3]. He was also site coordinator for an FP5 Research Training Network on visual modelling systems.

Runnalls' research focuses on terrain referenced navigation [Runnalls:1,3], Bayesian networks [Runnalls:2] and Markov Chain Monte Carlo methods. He has strong research collaborations with the navigation systems group of QinetiQ Ltd., Farnborough (formerly DERA). He has collaborated and consulted with them since 1989.

Plans

The interdisciplinary application of computing will remain the group's unifying principle.

Chue, Freitas and Johnson plan to build on their existing key strength in system biology and bioinformatics, by developing several new strong international collaborations. For instance, Freitas recently spent one month as a visiting researcher at the European Bioinformatics Institute (EBI), initiating collaboration with world leaders in bioinformatics, particularly Dr Rolf Apweiler.

Rodgers aims to integrate diagrammatic and textual notations to create highly expressive hybrid reasoning systems. In the layout area, he plans to test his visualization techniques by interdisciplinary work with geographers. Johnson and Rodgers plan to develop closer links with the TCS group.

Evidence of esteem

D.F.Chue

3

Guest editor for special issue of Artificial Life (11(4), 2005) on Dynamical Hierarchies.

4

Invited Speaker, Workshop on Biology and Physics, at University of Oslo, 2004.

5

Invited Speaker, Workshop on Genome and Protein Networks, National Centre for Theoretical Science, Taiwan, August 9, 2004.

A.A.Freitas

6

Program Committee Chair: 8th European Conference on Artificial Life (ECAL-2005).

7

Editorial Board Member of Intelligent Data Analysis: An International Journal.

8

Editorial Board Member of International Journal on Data Warehousing and Mining

9

Editorial Board Member of International Journal of Computational Intelligence and Applications

C.G.Johnson

10

Associate Editor, Journal of Artificial Evolution and Applications.

11

Editor: "The Society for the Study of Artificial Intelligence and Simulation of Behaviour" of AISB Quarterly, 2006-present.

12

Track Chair, EvoWorkshops conference, 2003 and 2004 and editor of LNCS proceedings.

J.C.RobertsCatB

13

Editorial Board Member of Information Visualization journal.

J.TimmisCatB

14

Initiator and co-chair of the first international conference on artificial immune systems (ICARIS-2002) at Kent.

15

NSF Advisor on bio-inspired computing, 2003.

Cognitive Systems

Primary staff: H.Bowman (head), R.A.Cowelle

Associated staff: C.G.Johnson

Statistics over 3 years: Income £102,685; 1 research degree awarded; 2.5 current students + 2 writing-up.

The Cognitive Systems group is the Computing Laboratory arm of the Centre for Cognitive Neuroscience and Cognitive Systems (CCNCS). The CCNCS was set up in 2004 to harness the potential for cross-disciplinary research at the junction of Cognitive Neuroscience and the Computational Sciences. It is well established that in order to make breakthroughs in understanding human cognition a broad spectrum of discipline-spanning techniques needs to be explored. Accordingly, the CCNCS undertakes research in which behavioural and neurophysiological experimentation informs and is informed by the construction of computational models, e.g. [Bowman:2][Cowell:1-2]. Such studies then enable the construction of artificial systems.

In the three years of its existence, the Centre has flourished and now boasts a dynamic seminar series and active members from a spectrum of departments, including Computing, Psychology, Electronics, Physical Sciences, Kent Institute of Medicine and Health Sciences, Philosophy and Film Studies (http://www.cs.kent.ac.uk/projects/cncs). At least 12 individuals from outside this UoA are regularly involved in its activities. The CCNCS recently opened an EEG recording facility, which is used for theoretical and applied research on the neural basis of cognition.

The group investigates a variety of cognitive phenomena. For example, we have developed computational theories of human emotion and attention [Bowman:1] and applied these in the context of human-computer interaction and affective computing. Perhaps most excitingly, we have demonstrated how EEG can be used to guide adaptive computer interfaces, which adjust the information they present according to the cognitive state of the user. We have also championed the application of formal methods (such as process algebra and model checking) in the context of computational modelling of cognition.

Success of the cognitive systems group is witnessed by the funding received, including the EPSRC project "Salience Sensitive Control in Humans and Artificial Systems" and an RCUK Academic Fellowship in "Cognitive Science and Robotics". Cowelle, a new appointee to this Academic Fellowship, strengthens these activities by bringing to the group her work [Cowell:1-4] on the modelling of rule extraction in category learning.

This group is strongly linked by cross-membership with the TCS and AII groups.

Plans

The group will expand the Brain Imaging activities, with particular emphasis on fMRI and EEG recording. Significant cross-fertilization is expected from the broadening and integration of Cowell'se work by exploiting these techniques. A key focus will be on the application of machine learning to the analysis of imaging data, thereby reaping the benefit of the unique cross disciplinary skill-set offered by the centre. We will apply our cognitive neuroscience findings in the context of cockpit and brain-computer interfaces.

Evidence of esteem

H.Bowman

16

Member of IFIP WG6.1, Architecture and Protocols for Computer Networks, throughout period.

17

Editorial Board of Springer journal New Generation Computing, throughout period.

18

Chaired and edited proceedings of 8th Neural Computation and Psychology Workshop (Connectionist Models of Cognition & Perception II), Aug 2003.

Computing Education

Primary staff: J.E.Carter, S.A.Fincher (head), M.I.Kölling, I.A.Utting

Associated staff: D.J.Barnes

Income £366,329; 3 research degrees awarded; 2.25 current students.

The Computing Education (CompEd) group are actively engaged in building capacity for computing education research as a distinct endeavour [Fincher:2] and in establishing research-only publication venues; to this end, the group hosted the second ACM International Computing Education Research (ICER) conference in 2006. Creating this specialist event has given a research focus lacking in long-established conferences (e.g. IEEE Frontiers in Education; ACM SIGCSE Symposium) which cover all aspects of educational practice, marginalizing research.

CompEd researchers distinguish Discipline-specific education research activity from scholarly examination of their practice. Our expectation for rigorous research is that studies should involve more than one classroom (by multi-institutional involvement or by multiple instantiations of an intervention) to demonstrate generalisability.

CompEd is a young and growing community, without established paradigmatic research norms; thus various methodological traditions are used, from statistical investigations [Carter:3-4] through empirical studies [D.Barnes:2,4] to interpretative [Fincher:1] and theoretical contributions [Fincher:2].

A special strength of the CompEd group is our work on BlueJ, an initial learning environment for object-oriented programming, using Java. BlueJ is widely used in University-level teaching (865 universities worldwide are using BlueJ (http://www.bluej.org/about/users.html)). Kölling is lead designer, responsible for its creation and development [Kölling:4]. Recent work has integrated BlueJ with a professional environment (NetBeans), to ease the student's transition to industrial-strength techniques [Utting:3]. This work contributed to the Laboratory's recognition as a Sun Microsystems Centre of Excellence in Object-oriented Education (http://www.cs.kent.ac.uk/coe/).

Kölling's recent work targets pre-University learning through development of the visualization environment Greenfoot, leading to greater engagement by school children with Computer Science [Kölling:2]. Concurrently, Utting has worked on the BlueJ extensions framework, increasing the environment's utility to different teaching approaches at the tertiary level [Utting:1], and on the ACM Java Libraries [Utting:2].

Another strong CompEd contribution is in the development and execution of multi-institutional (often multi-national) research studies: Utting worked on one of the pioneering projects [Utting:4] and Fincher subsequently refined and strengthened the methodology [Fincher:3]. We have worked on studies with participants drawn from over 20 institutions [Carter:1-2][Fincher:2,4][Utting:4].

Plans

The CompEd group aims to extend work based on the instrumentation of teaching environments and the use of these tools to mount new multi-cohort, multi-institution studies of CS student learning. Plans include more detailed, in-depth work on the Greenfoot environment and its support for teaching and learning computer science. The objective is to provide a beacon for CompEd research, based on Kent tools, which enable stronger quantitative tests of our theoretical contributions than would be possible in a purely theoretical group. We intend working with a regional network of high schools, drawing on our existing expertise in the design of multi-institutional studies.

Evidence of esteem

S.A.Fincher

19

Editor-in-Chief of journal Computer Science Education.

20

Recipient of the 2003 IEEE Computer Society Computer Science & Engineering Undergraduate Teaching Award "for sustained contributions to undergraduate computer science education through rigorous examination of teaching effectiveness and fostering and promoting research in computer science education".

21

Chair: Innovation and Technology in Computer Science Education (ITiCSE) 2001.

22

Invited keynote address, Australasian Computing Education Conference 2004

I.A.Utting

23

Programme Co-Chair, ACM Innovation and Technology in Computer Science (ITiCSE), Denmark, 2002.

Information Systems Security

Primary staff: D.W.Chadwick (head), H.Ragab-Hassene

Associated staff: A.M.King, R.DeLemos, P.F.Linington

Statistics over 2.5 years: Income £340,010 (+£205,416 JISC-funded research); 1 current student + 2 writing-up.

The ISS group's mission is to research and develop user-friendly security solutions for distributed systems. Security systems that are hard to use are likely to be ignored or bypassed, leading to compromise. Major research areas include: public key authentication infrastructures (PKI) [Chadwick:2], privilege management (authorization) infrastructures (PMI), trust management systems and user privacy. The recent appointment of Hassen adds an application-oriented key management perspective [Hassen:1-4]. The group has attracted 21 research grants since 2001, leading to 80 publications.

PKI research has included case studies in the healthcare domain, defining best practice in operating a PKI system and determining deficiencies in current PKI technologies. User privacy research has focussed on privacy protection of electronic prescriptions and in trust management.

PMI-based research has increased dramatically, being a key component of Federated Identity Management; nine research projects have been completed, with six journal papers and over 20 conference/book chapters published. The keystone of this has been the EC PERMIS project, which started in December 2000, and culminated in the first X.509 PMI-based system being built and piloted in Barcelona, Bologna and Salford. The PERMIS software [Chadwick:3] was released as open source, distributed via the US NMI software initiative, and via its integration into the Globus Toolkit, OMII (UK) and Shibboleth, and now into VOMS and the UK National Grid Service. Subsequent PMI-related projects have specified open protocols to interface grid applications to PMI systems, adding more sophisticated security features to PERMIS, such as separation of duties, delegation of authority, coordinated decision making, and aggregate attributes from independent identity providers.

PMIs and PKIs are types of trust management systems, but they traditionally depend on binary values of trust for certificate issuers. Scalar values of trustworthiness will provide finer granularity of control [Chadwick:4]. Chadwick has researched various aspects of this, including the trustworthiness of Certificate Authorities, distributed systems and reputation systems, contributed to the standards making process and will next contribute to computing the level of assurance of authentication.

Chadwick has worked with Prof Sasse of UCL on two projects focusing on ease of use of PMI policy creation [Chadwick:1]. Usability has also been studied in the context of electronic prescribing (with Hope Hospital) and delegation of authority (with Sinnott at Glasgow).

The associated work on other aspects of security is linked to the Systems Architecture group and takes a longer term view. King applies abstract interpretation to security analysis, de Lemos investigates peer-to-peer security architectures, and Linington uses model-driven approaches to manage system security.

Plans

We will explore usability, privacy and policy-based authorization issues. Hassen's involvement will complement the user-based emphasis with content-specific considerations. The goal is to achieve distributed decision making, involving heterogeneous policy languages and authors. The broad PERMIS user base and its integration into the National Grid Service will provide fertile ground for this, as will Chadwick's close links with the health sector. Chadwick's recently awarded €900,000 FP7 grant "Trusted Architecture for Securely Shared Services" will explore these directions in the health and employability domains.

Evidence of esteem

D.W.Chadwick

24

Organiser and PC chair of 2004 IFIP CMS conference (Communications and Multi-media Security).

25

Keynote speech at First EuroPKI Workshop, Greece, 2004.

26

PC Chair for Second EuroPKI Workshop, 2005.

27

PC Chair for TERENA (Trans-European Research and Education Networking Association) Conference, TNC2004.

Systems Architecture

Primary staff: D.J.Barnes, F.R.M.Barnese, R.de Lemos, T.R.Hopkins, R.E.Jones (head), P.F.Linington, N.S.Ryan, P.H.Welch

Associated staff: M.I.Kölling

Income £1,017,426; 23 research degrees awarded; 13.75 current students + 2 writing-up.

The group focuses on the specification, design and implementation of computer systems, from the very large to the very small, being characterized by the interplay between theory and practice. It has five main research themes: modelling large-scale distributed systems; mobile systems; automatic memory management ('garbage collection'); concurrency; and software dependability and testing. The practical orientation of the group has resulted in software tools to support the development, execution, testing and debugging of software systems, and the teaching of professional systems development concepts and techniques.

There is a natural synergy between academic research and industry; international and industrial collaborations therefore underpin the group's work [F.Barnes:1][DeLemos:4][Linington:3][Jones:3,4][Hopkins:1]. Members are active in promoting research (e.g. in International SPACE workshop, EPSRC top-rated MMnet network, European InterOp-NoE, etc.).

The Laboratory has a long-established reputation for research into distributed systems, exemplified by the work by Linington on enterprise-scale systems, de Lemos on dependable systems, and Ryan on mobile and ubiquitous systems.

Linington's work on modelling enterprise policies and contracts has influenced standards for Open Distributed Processing, particularly their conceptual framework for roles and viewpoints [Linington:2-4]. His work on Model Driven Engineering has evolved from the use of transformations to extract performance models at the start of the period to aspect weaving [Linington:1].

de Lemos's work on dependable systems established a new approach to the verification, validation and synthesis of dependable software [deLemos:2]. He has analysed failures [deLemos:4] and rigorously studied exception handling [deLemos:3] in dependable, component-based systems. Funded by NCR, he has applied work on artificial immune systems to anomaly detection in ATMs [deLemos:1].

Ryan is the principal architect of the MobiComp infrastructure for context-aware mobile systems [Ryan:1]. With AII members, he has applied techniques from artificial immune systems to recognizing familiar contexts [Ryan:3]. His MobiComp infrastructure supports cultural heritage applications through the FieldMap data collection and access tools [Ryan:2,4].

Jones's research concentrates on efficient garbage collection. He applies theory to practice, in order to improve the quality and performance of modern garbage collectors. He has formalized distributed garbage collection algorithms in an implementation-independent manner, discovering omissions in the published algorithm widely used by Java-RMI, and proving corrections [Jones:2]. He studies the lifetime demographics of objects in order to focus garbage collection effort most effectively. This work, supported by three IBM Faculty Awards, led to the Beltway GC framework, developed collaboratively with the Australian National University and the Universities of Texas and Massachusetts [Jones:4], and to a novel scheme for deriving programmer intention in order to better manage Java objects [Jones:1]. Debugging garbage collectors is hard. Jones's widely adopted GCspy visualization framework [Jones:3], developed in collaboration with Sun Microsystems, provides the GC developer with an intuitive, scalable and high performance graphical view of their collector.

Welch and F.Barnese focus on abstractions and implementations of concurrency in programming languages. They are the principal designers of the occam-pi language, which preserves the elegance and power of the occam concurrent programming language yet adds dynamic asynchronous multiplexed communication from the pi-calculus, essential for the construction of modern software systems [Welch:3,4]. They show how concurrency abstractions can be built into languages ensuring correctness (freedom from deadlock, livelock, race-hazard and aliasing errors) yet still achieve very high performance. F.Barnese focuses on implementation of occam-pi [F.Barnes:2]; the effectiveness of their design strategy of numerous fine-grained processes [F.Barnes:3] underpins the collaborative EPSRC TUNA project [Welch:1] and the lightweight, multi-threading run-time operating system kernel of F.Barnes'e EPSRC RMoX project [F.Barnes:1]. They are exploring design patterns and design tools for encouraging safe parallel programming in their new £0.5M EPSRC CoSMoS project.

The correctness of many software systems can only be established through rigorous testing. Hopkins concentrates on using testing to improve the quality of the numerical software that underpins most scientific modelling applications [Hopkins:2,3,4]. He focuses on the implementation of parallel numerical algorithms to solve partial differential equations describing industrial processes [Hopkins:1]. D.Barnes develops tools and techniques to support open, collaborative testing of widely-used software libraries [D.Barnes:1,3], aiming both to demonstrate improved quality through error detection, and to provide open evidence of test coverage. Kölling has worked on tool support for testing distributed systems [Kölling:1]. His novel tools are widely used to inculcate testing of programs early in a programmer's career, particularly in education [Kölling:3]. Starting a systematic testing practice at this early stage has been shown to influence programmers' attitudes to the use of professional testing tools later in their career.

Plans

The Systems Architecture group will continue to focus its efforts on tools and techniques for the development of reliable and efficient systems. The EPSRC RMoX and CoSMoS projects will allow F.Barnese and Welch to apply their communicating processes approach to the design of operating systems and complex systems. The EPSRC LACE project will allow Jones and King to extend the Beltway work to yet more flexible GC schemes better tailored to individual program behaviour by exploiting static and other off-line analyses. D.Barnes's and Hopkins's aim to produce a toolset with capabilities allowing them to extend their methodologies and techniques, testing a far larger body of numerical and scientific codes. BSSI (Norway) is funding Hopkins to investigate efficient usage of multicore processors in numerical industrial-engineering problems.

Evidence of esteem

F.R.M.Barnese

28

Best paper awards, Communicating Process Architectures conference, [F.Barnes:1][F.Barnes:2].

R.de Lemos

29

Co-chair Latin American Symposium on Dependable Computing (LADC) 2003 and Programme Committee LADC 2005.

T.R.Hopkins

30

Algorithms Editor of Association of Computing Machinery: Transactions on Mathematical Software (ACM TOMS).

31

Member of IFIP WG.2.5 on Numerical Software, elected 2003.

R.E.Jones

32

Distinguished Scientist, Association for Computing Machinery (ACM) 2006.

33

Keynote Speaker, International Lisp Conference, Cambridge, 2007.

34

Three IBM Faculty Awards (2002-2004).

35

Honorary Fellow, University of Glasgow (2005).

S.J.H.KentCatB

36

Royal Society Industry Fellow, 2002.

P.F.Linington

37

Active participant in standardization via ISO/IEC/JTC1/SC7/WG19; most recently Principal UK Expert to Kyoto meeting, 2006.

38

Member of International review of Simula Research Laboratory 2004; coordinated report to Norwegian Research Council, ISBN 82-12-02017-7, laid before Norwegian Parliament and referenced in debate.

39

Member of International Research Advisory Board for Australian CRC for Enterprise Distributed Systems (previously DSTC), 2001, 2004.

40

Task group leader (TG7), EU InterOp Network of Excellence, 2005-2007.

I.W.MarshallCatB

41

Royal Society Industry Fellow, 2001.

42

Chair IT, Electronics and Communications Group (ITEC), Institute of Physics

N.S.Ryan

43

Programme Committee Co-Chair: 6th International Symposium on Virtual Reality, Archaeology and Cultural Heritage (VAST2005), Pisa, 2005.

44

Steering Committee Chair and PC member: International Conference on Computer Applications and Quantitative Methods in Archaeology.

45

Invited lecture, 24th International Research Symposium, International Research Centre for Japanese Studies Kyoto, Feb 2005

P.H.Welch

46

Member of IFIP WG 2.4: System Implementation Technologies, elected 2002.

47

Programme Committee Chair, Communicating Process Architectures CPA 2001-2006.

48

Vice-Chair (Models, Methods and Programming Languages), EuroPar 2003 and 2006.

49

Conference Co-Chair, CPA-2000, CPA-2007

Theoretical Computer Science

Primary staff: E.A.Boiten, O.Chitile, R.S.Gomeze, S.M.Kahrs, A.M.King (head), S.J.Thompson

Associated staff: H.Bowman

Income £1,273,087; 12 research degrees awarded; 3.5 current students + 3 writing-up.

The TCS group draws on foundational ideas in computer science and applies them to practical problems in programming languages, program development, verification and formal methods. It does this by developing tools, such as refactorers and tracers, to underpin the program development cycle. It also investigates diagnostic techniques such as abstract interpretation and model checking [Bowman:3]. As well as investigating the correctness of existing programs, it is developing refinement techniques that systematically flesh out a specification to obtain implementations correct by construction.

Verification: Specifications need to be abstract, referring to idealized mathematical notions such as real numbers. Implementations take into account resource bounds, and cannot be expressed at the desired level of abstraction of initial specifications. Boiten bridges this gap through program development using refinement rules [Boiten:2,3] to construct sequences, and even take limits, of increasingly more correct specifications [Boiten:1]. Closely related to this is work on viewpoint consistency [Boiten:4][Bowman:4].

King applies refinement's dual, namely abstraction. For logic programs he has developed abstraction techniques [King:4] that infer classes of inputs for which a program conforms to its specification. Thompson and Bowman reason about temporal properties of systems, specifically studying interval temporal logics (ITL) with the projection operator and showing, among other things, how to axiomatise logical consequence within this system [Thompson:4].

Program transformation: For Haskell programs, Thompson has developed the Hare refactoring tool [Thompson:3]; refactoring involves changing the design of a system without changing its semantics, thus facilitating program reuse and generalization. King has developed transformation schemes that facilitate the reuse and generalization of logic programs [King:1]; these transformations preserve the logical semantics but yield programs that are operationally more general. Chitil has applied program transformation in program development tools themselves to construct portable tracing tools for functional programs [Chitil:3].

In the world of rewriting, there is particular interest in verifying properties such as termination and confluence. In the specific case of infinitary rewriting, termination is replaced with convergence, and Kahrs has developed foundational proof techniques which show that this property holds for various term metrics [Kahrs:1].

Principled implementation: Types are useful for diagnosing errors in programs and are therefore an important aspect of program development; Chitil has developed methods for locating the cause of type errors in type-inference-based programming languages [Chitil:4]. Kahrs has explored the constraints that typing put on expressibility and shown how untyped, but semantically sound, programs expressed in PCF are equivalent to well-typed programs [Kahrs:3].

Chitil has developed algorithms that take advantage of the unique facilities offered by declarative languages, showing that the pretty printing problem can be solved elegantly by exploiting the cyclic definitions supported by lazily evaluated languages [Chitil:1]. Kahrs has shown how tree balancing invariants can be realized merely using the type system of Haskell to shift the cost enforcing balance to compile-time [Kahrs:4]. King has demonstrated how compile-time analysis techniques can be implemented efficiently, using logical variables and synchronization devices provided by logic languages [King:1-3].

Application of theory: The utility of theory is tested by its application to real problems; the group always aims to prove its outputs in the real world. Examples are the work of Thompson on refactoring [Thompson:3], application of model theory to problems in visual languages [Thompson:1] and the transfer of functional idioms to the XML description of reactive systems [Thompson:2].

Plans

Over the next five years, the TCS group will focus its efforts on reasoning, checking and verifying salient properties of programs, for instance, by devising termination criteria for term-rewriting systems, developing tracing tools for functional programs, and applying model-checkers to real-time and imperative programs. Already, the recently awarded €404,000 ProTest FP7 project will let Thompson develop fault-finding tools for Erlang programs so as to improve the reliability of service infrastructure software, and an upcoming EPSRC-funded industrial secondment (EP/F012896) will allow King to apply model-checking to discover security vulnerabilities in C code.

Evidence of esteem

E.A.Boiten

50

Guest editor: four special issues of Formal Aspects of Computing.

51

Program chair: Mathematics of Program Construction 2002 and program committee membership throughout period.

52

Program chair: Integrated Formal Methods 2004 and PC member 2005, 2007.

53

Guest editor: special issue of Science of Computer Programming.

A.CavalcantiCatB

54

Royal Society Industry Fellow 2003

O.Chitile

55

Advisory Committee Member: International Symposium on Implementation and Application of Functional Languages, IFL.

J.DerrickCatB

56

Vice-chair IFIP WG6.1 (Architectures and Protocols for Distributed Systems).

57

IFIP Steering Committee member for conferences FMOODS; FORTE/PSTV; TESTCOM

R.S.Gomeze

58

EPSRC Fellowship in Theoretical Computer Science 2006-2009

A.M.King

59

Associate Editor: Higher-Order and Symbolic Computation.

60

Editorial Advisor: journal Theory and Practice of Logic Programming (formally Journal of Logic Programming).

S.J.Thompson

61

Editor, Book reviews, Journal of Functional Programming.

62

Chair: ACM Special Interest Group on Programming Languages (SIGPLAN) workshops in 'Functional and Declarative Programming in Education' 2002 and 2005.

63

Invited speaker: 5th International Summer School on Advanced Functional Programming (AFP 2004), 2004.

64

Member of the International Scientific Committee of Institut de Recherche sur les Composants Logiciels et Matériels pour l'Information et la Communication Avancée (IRCICA) Lille, France, 2005-8.

J.C.P.WoodcockCatB

65

Chair of Grand Challenge 6: Dependable Systems Evolution.

Administrative infrastructure

Good administrative support is essential in making sustainable and effective use of our research resources. The University has a well provisioned and supportive central research office, which we augment within the laboratory to give researchers rapid and continuous support in managing their projects. We have, for eight years, provided a full-time research administrator, who promulgates funding opportunities, supports the grant application process and assists in managing projects. The research administrator also organizes postgraduate research training events, such as induction activities and a poster fair for second-year students. They take the lead in publicizing activities on the web and via a research newsletter. This research support complements the normal support for staff by the departmental general and financial administrators.

Technical Infrastructure

The ubiquitous technical computing infrastructure for the laboratory is a balance between large Unix servers and regularly updated desktop and laptop facilities. All staff benefit from a rolling replacement programme, ensuring no personal machine is more than three years old, and all research postgraduates are provided with a new computer on arrival. There are also more specialized resources. There is a flexible networking test bed (with dedicated routers, switches and servers, for constructing test configurations) and, for larger-scale experiments, the Laboratory is a member of PlanetLab. For robotics, there is a laboratory with platforms ranging from a large Pioneer DX3 with many sensors down to a substantial holding of experimental kits. For parallel processing, there is a 32-processor cluster, primarily associated with Welch's TUNA project. Chu has recently been awarded two MacPro dual-quad-core systems for his computational biology work. All these facilities are integrated by fixed and wireless network provision.

This departmental technical core is supported by a five-person technical development team who, in addition to their system management responsibilities, support short-term research initiatives, or projects in the start-up phase before funding and dedicated research staff are in place. This team is widely known for the high quality of their support for the UK Mirror Service. The campuses are connected by a high capacity network infrastructure and by gigabit access to the region and the SuperJanet core.

The University Library holds a broad collection of books, conference proceedings and journals. The emphasis is on access in electronic form, accounting for 60% of CS library spend this year (£71k).

Human resources

Early career researchers All newly appointed academic staff and Research Fellows undertake the Postgraduate Certificate in Higher Education (PGCHE), equipping them for the teaching and research aspects of their future careers. The PGCHE covers techniques for research and research management and provides a forum for new researchers to interact with peers and so achieve a broader understanding of the approaches prevalent in different disciplines. This has stimulated interdisciplinary collaborations within the University, e.g. with staff in Biosciences.

The laboratory has a policy of limiting the load placed on new staff, giving them the headroom necessary to develop their careers. This is not just to attend the PGCHE, but provides a generous allowance of time for development of research ideas and preparing funding bids to sustain their research in the longer term. Each appointee has a mentor who assists and advises them in formulating their plans. The head of the research group concerned monitors this process and provides more strategic advice. New staff are given priority when departmental research bursaries are allocated, so that they gain experience of supervision as soon as possible.

Staff loading In the Laboratory's load model, staff are given a substantial baseline time allocation (450hrs) to pursue their ideas; they supplement this by attracting additional resources or producing outputs and funding bids. All staff are given a basic budget of £500/year to cover some travel and smaller items of equipment, supplemented by bids against the general travel and equipment funds. Research-active staff are strongly encouraged to take study leave (one term after seven) to gain access to new ideas and foster external collaborations.

Postgraduate studies

The Laboratory maintains a strong postgraduate research culture, and has allocated substantial funding over the last decade to provide bursaries, ranging from payment of fees to full support. The Laboratory has an earmarked bursary fund, and has funded 48 of the 127 students admitted since 2001.

Research postgraduates participate in a broad range of career development and professional skills training. Departmental training provides subject-specific skills, complemented by an extensive portfolio of Robert's-funded provision at the Faculty level. Students are encouraged to engage in national or international events such as summer schools and solicit feedback on their ideas while in a formative stage. In the TCS area, Kent is the leading member of the TCS-SOUP (Theoretical Computer Science for SOuth-eastern Universities PhD students) postgraduate training consortium. Students have a budget of up to £500 per year to support conference travel, and are encouraged to publish regularly as soon as results allow. Beyond the normal academic induction process, new students engage in social team-building events, such as an outward-bound weekend, to strengthen links and encourage mutual support within each cohort.

In addition to their supervisor (or supervisors, co-supervision being common in interdisciplinary areas), each student is supported by a supervisory panel; two further academics join the supervisor in offering advice and reviewing the student's work at key points. Panel members assess the mini-thesis, a mandatory progression hurdle after the first year of research. The panel is an essential element in the monitoring process from induction to submission, making a formal report on progress every nine months.

Research Management

Group activities are coordinated via termly meetings of a Research and Enterprise Executive. This is the main internal management tool, where group heads meet to contribute to planning and ensure the smooth running of the whole research process. The research management processes are lightweight to maximize research productivity, reflecting the University's commitment in reviewing its procedures to reduce administrative overheads.

Individual researchers meet with the Department Head and Research Director annually to discuss plans and ensure that individual goals are ambitious, yet achievable, opportunities for collaboration are followed up and resources are deployed effectively. To maintain coherence of the department's research activities, research elements are included in the regular laboratory staff meetings; these cover opportunities, changes of process and showcase spots to raise general awareness of individual project activity.

The department produces annual updates to its research plans, which are reviewed by, and inform, planning at the Faculty and University levels. The Laboratory uses its pool of industrial contacts as a sounding board for research plans. The range of issues is too diverse for a single industrial panel to cover; invited groups from our contacts are selected to discuss particular strategic issues, ensuring continued relevance of research.