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.
If you would like to give a seminar to the department in future or if you need more information, please contact the seminar organiser Ken Turner (Phone 01786-467-420, Email kjt@cs.stir.ac.uk).
Prof. Al Dearle, University of Stirling
The computational paradigm with which we are all familiar has remained fixed for the last 50 years: we are used to running programs on a single fixed piece of hardware and user interaction is through a single fixed display device. Fuelled by the ubiquitous nature of the internet, increasing hardware power, and availability of high speed networks, a radical paradigm shift to a new computational model is now possible. In this new paradigm processes are not tied to running on a single machine and interacting with users via a single display device. Instead, they are free to migrate with users providing truly ubiquitous computing environments. However, there are a number of technical problems that must be solved in order to accommodate this model. This talk investigates these technical issues and examines some systems that are addressing them.
Dr. Steve Marsh, University of Stirling
The ACORN architecture is a multi-agent based system which leverages Stanley Milgram's `Small World' theory to provide directed filtering and dissemination of information in communities of human users. A key idea is that of `information as agent' in which every piece of information (part of, or whole, document, image, sound, query, etc.) is embodied as a (semi-)autonomous agent in the system which has a `knowledge' of the community of humans and agents it exists in. It uses this to route itself to potentially relevant human readers. A component of the system is the ability for information agents to be able to communicate with each other (in special places) in order to better achieve distribution of information without the need for human intervention. We believe that ACORN can help in managing information flow between people. The presentation will discuss the architecture, its workings, and the theories behind it. It will also present the working prototype of the ACORN Messaging System, and ideas for further work (notably the ACORN Web System).
Prof. Ken Turner, University of Stirling
The informal requirements for the invoicing case study will be analysed and interpreted. This will lead to a high-level specification architecture that can be formalised. Specifications will be presented in LOTOS (Language Of Temporal Ordering Specification). For comparison, specifications will also be presented using E-LOTOS (Enhancements to LOTOS) - the new version of LOTOS currently being standardised. Since LOTOS allows a balance to be struck between process-oriented and data-oriented modelling, specifications in both styles will be given.
Dr. Tom Melham, University of Glasgow
This talk will sketch the main ideas behind a project at the University of Glasgow to develop new methods for realising circuits as hardware which is modified at execution time and whose correctness is formally assured. The project is based on a novel combination of the technology of Field Programmable Gate Arrays and the functional programming technique of Partial Evaluation. I shall aim to be fairly non-technical, concentrating on the overall vision of the project, but I will also try to give an idea of the combination of hardware specification/verification and reasoning about algorithms we propose to use.
Dr. Simon Jones, University of Stirling
Traditionally, purely functional programming languages have had a very poor popular image: they were criticised, justifiably, for their profligate use of computer memory, their poor execution speed, and their abysmal input/output capabilities. Given these difficulties, their virtues for programmer efficiency did not hold sway, and they had little impact on the world of real software development. However, more recently, there have been developments in purely functional programming languages that see them well on their way to dismissing the criticisms above. In this talk I will survey the history of purely functional programming languages, and the sound reasons for their former poor characteristics. I will explain the recent developments that have led to improvements, and discuss their impact on the nature of the functional programming paradigm itself.
Graham Birtwistle, University of Leeds
Asynchronous hardware design is attracting renewed research and industrial interest because of its compositional properties and potential for low power consumption (much of today's hardware is battery driven and asynchronous circuits only do work when there is work to do). The talk will present work nigh completed on designing, specifying and verifying the major subsystems of an AMULET-like micro in CCS. AMULET is an (already working) asynchronous version of the ARM designed by Steve Furber's team at Manchester University. Like AMULET1 our micro is 2-phase, but like StrongARM, we have separate instruction memories, and like AMULET3, a wide arithmetic pipeline for greater parallelism. By following an obvious (and proven) design discipline, and making use of old theorems on the register bank (from previous work on AMULET1) and some obvious recent ones on pipelines, we have reduced the state minimisation problem by a handsome factor and the whole model fits comfortably on the Edinburgh CWB. The talk will cover the basic architecture, enough on CCS and asynchronous architecture to get you by, and then detail the 3 major system blocks (fetch unit, arithmetic pipeline, and save unit) at the RT level, and how the design deadlocks (er, flaws) were fettled. The work is being carried out with Matt Morley and Chris Tofts at Leeds University with plenty of help from Jim Garside at Manchester.
Up one level to Computing Science Seminars