The Department of Computing Science and Mathematics presents the following seminars. Unless otherwise stated, seminars will take place in Room 4B94 of the Cottrell Building, University of Stirling from 15.00 to 16.00 on Friday afternoons during semester time.
If you would like to give a seminar to the department in future or if you need more information, please contact the seminar organiser Carron Shankland (Phone 01786-467-444, Email ces@cs.stir.ac.uk).
Ian Wakeman, School of Cognitive and Computing Science, University of Sussex
The communications networks of the future (and many networks of today) will not send requests directly from a source to a destination. Instead, each request will be subject to various levels of indirection, where a request is ``intercepted'' by some computational device. The device will use the request as input to some program, which may satisfy the request, generate further requests onwards or drop the request.
In this talk I'll discuss some of the issues raised by a programmable network in which each request may be subject to multiple levels of indirection, illustrating problems and solutions using recent work from the networking group at the University of Sussex on programming languages and algorithms.
Brian Ross, Department of Computing Science, University of Glasgow
The ISO semantics for Full LOTOS generates infinitely large labelled transition transition systems (LTSs) for a large class of specifications.
The symbolic semantics of Calder & Shankland removes the infinite branching aspect of these processes by providing a representation in terms of symbolic transition systems (STSs). A suitable notion of symbolic bisimulation over these STSs has also been defined.
Standard bisimulation algorithms cannot be applied to STSs as the underlying systems are infinite. I present an algorithm for symbolic bisimulation which reduces the problem of checking a potentially infinite system to that of reasoning about (universally quantified) boolean expressions. This is an extension to the similar work for value-passing CCS of Hennessy and Lin.
I will outline the soundness and completeness proofs required by such an algorithm and shall demonstrate a prototype if time allows.
John Zeleznikow, Joseph Bell Centre for Forensic Statistics and Legal Reasoning, Faculty of Law, University of Edinburgh
The use of knowledge discovery techniques to construct intelligent legal decision support systems in discretionary domains will
1) enhance consistent decision-making;
2) by making legal decision making transparent lead to increased
confidence in the justice system ; and
3) provide support for alternative dispute resolution
The development of legal decision support systems on the World Wide Web will increase access to legal knowledge. In this talk we discuss how to place legal decision support systems on the World Wide Web. We discuss the role of argumentation and negotiation in building such systems. We describe two tools (WebShell and ArgumentDeveloper) we have developed for placing legal decision support systems on the World Wide Web.
Jill Ireland, University of Stirling
The multifocal ERG can evaluate the retina at up to 103 locations. It describes the way in which the response of the retina changes its characteristics after it receives a flash of light, and this response contains significant diagnostic information.
The use of finite field theory to generate primitive polynomials and zech logarithm analysis is able to predict which primitive polynomials are suitable for m-sequence generation for multifocal electroretinography. This has important implications for world-wide developers of multifocal electrophysiology systems.
Jane Hillston, School of Computer Science, University of Edinburgh
Over the last ten years Stochastic Process Algebras have emerged as a innovative approach to the development of performance models of computer systems. In this talk, after a brief overview of the field, I will describe the development of one of these languages, PEPA, in some detail. In particular I will emphasise the benefits which can be gained from the use of a process algebra for constructing, manipulating and solving Markov process-based performance models. At the end of the talk I will discuss recent work which aims to enhance the expressiveness of PEPA towards modelling systems with mobility.
David Cairns, University of Stirling
This seminar will provide an overview of Microsoft's .NET initiative and will describe some of the core features of the C# language. The aim of this talk is to raise an awareness of .NET and its likely implications for web based software development rather than provide an in-depth review of .NET and C# features.
The main topics that will be discussed are:
Sharon Curtis, University of Stirling
This talk presents several algorithms to solve a problem posed by a programmer wanting to allocate batches of drugs for a double-blind medical trial. One of the algorithms is very simple, and makes use of laziness to achieve a linear time complexity. However, the proof of its efficiency is far from simple, and involves the largest multinomial frequency distribution.
Julie Hodgkin, University of Stirling
The talk will focus on a case study conducted last year in conjunction with Glasgow Decision Support Project (GDSP). This project was initiated involving both the Scottish Enterprise Group (SEG) and Strathclyde University. The project's objective was to help organisations in the Glasgow area make key business decisions by providing access to leading edge decision support methods and expert advice on the process.
The case study conducted revolves around one of the projects undertaken by the GDSP which involved aiding a Glasgow-based charity concerned with the care of disabled clients. The organisation at the time of the study had a residential unit for those patients requiring 24-hour care, and a day-centre. Broadly, the issue facing the organisation was how these facilities should be developed in the future.
Six workshops were run with the client, during which the facilitators of the workshops used three DSS to aid in the decision making process.
In the presentation I will give a detailed description of the case study, outlining the various stages of the decision making process, and how the software developed aided in this process. In addition, I will discuss the various outcomes and learning points that arose from the study.
Portia File, Abertay University
We are designing a communication aid for people who are literate but who cannot speak. Our potential users can type in phases to speak during a conversation but, even with word predication support, this is far too slow for a satisfactory interaction.
The problem for us is how to help our users participate in conversation satisfactorily and we have designed our aid based on user-constructed pre-stored speech. As interface designers we must help them quickly find appropriate things to say. Our retrieval strategy is based on pragmatic features of conversation and we provide evidence that these features do support user goals. The system is currently being re-designed to enhance the user's social interactions and employment prospects by improving the system usability and by providing structured access to standard office communications.
Richard Bland, University of Stirling
This seminar will review the basics of XML and then give an evaluation of recent developments in XML processing. The talk will cover the following points: - A brief introduction to XML: what it is and why it's useful - Reading/writing XML in Java and Perl - A brief introduction to XSLT - XSLT processing in Java - XML/XSLT Web publishing frameworks, particularly those (like COCOON) based on Java servlet technology
Dave Budgen, Keele University
I have been interested in the topic of Software Design since the late 1970's. In my talk, I will explain a little about how I originally became interested in the topic, and how my ideas have evolved. I will review some of the work we have done in constructing `proof of concept' design support environments, and how the experiences of evaluating these has led to an interest in the more empirical aspects of Software Engineering. Finally, I will talk about some recent work in which we have studied the activities involved in designing component-based systems, and briefly review some other related ideas and developments, including our recently awarded EPSRC project to build a proof-of-concept information broker for the healthcare domain.
Hannah Harris, Birmingham University
We formalise a relationship between two previously unconnected areas: theory change and feature interaction. This will provide an interesting new application area for the logic of theory change, and a theoretical underpinning for the feature interaction problem which has a largely practical basis.
Update is an operation of theory change which is closely related to belief revision. The principal difference lies in the fact that belief revision models changing beliefs about a static world whereas update models a changing world.
A feature is a unit of functionality which extends or modifies the behaviour of the system into which it is integrated. The feature interaction problem arises when two or more features interact, causing the system to exhibit unexpected, and often undesirable, behaviour. Many approaches to feature integration and interaction detection have been proposed. In this research we use a feature construct for the model checker SMV.
We show that there is a strong connection between update and feature integration by, preliminarily, formulating SMV and the feature construct in propositional logic. We then go on to prove that the eight rationality postulates for update hold in the context of this theoretical formulation.
Evgeny Selensky, University of Glasgow
The long term goal of our research is to try to understand why the seemingly similar vehicle routing and job shop scheduling problems are dealt with in practice using different solution techniques.
In this talk, I will first recap on the notions of the vehicle routing and job shop scheduling problems, their differences and similarities. I will also describe the way we can represent these problems as graphs. The talk will focus on some specific graph transformations that, we hope, will allow us to answer the mentioned long term question.
I will present some preliminary results we have obtained using the graph transformations and ouline the future research directions.
Autumn 2001 Seminars
To
Computing Science and Mathematics Seminars (Current Series)
To
Computing Science and Mathematics Research
To
Computing Science and Mathematics Home Page