🚀 go-pugleaf

           University of Kent Computing Laboratory
           =======================================

            Advanced Training in Computer Science

The Computing Laboratory has available funding for students
attending a one year advanced training course in Computer Science
leading to an MSc (by research).  Applicants should have a strong
computing background.  The course consists of the following components:

+ Advanced Project - a supervised project leading to an MSc thesis.
+ Information Gathering and Management,
+ Selected Topics in Computer Science,
+ Computing Service training courses,
+ Personal Skills at Work,
+ Directed reading (set by supervisor).

Students on the programme will be allocated a supervisor who will
direct and coordinate their work.  Each student's programme
will be tailored to their background and the requirements of their
project.

For general information contact:

Steve Hill (sah@ukc.ac.uk) or David Morse (drm@ukc.ac.uk)
Computing Laboratory,
University of Kent,
Canterbury
Kent.
CT2 7NF

PROJECT LIST

The projects may be drawn from (but are not restricted to) the list
below.  The laboratory also has research interests in the following areas:

	+ Networks and Distributed Systems
	+ Parallel Processing
	+ Software Tools and Electronic Publishing
	+ Functional Programming and Theoretical Computer Science
	+ Database Technology and Object Management
	+ Artificial Intelligence and Expert Systems
	+ Numerical Computation, Simulation and Visualization


                     Formal Modelling of Distributed Systems
                     =======================================

An investigation into the use of formal description techniques, such
as LOTOS, Z and Raise, within the area of Open Distributed Processing.
We expect the student to follow one of our research initiatives in this
area.  There are a number of possibilities, including:

1. Specification and validation of distributed multimedia systems.

2. Development of techniques for consistency checking between ODP 
specifications.

3. Investigate techniques for replacing the data-typing part of the LOTOS 
specification language with a functional language, eg Miranda.

A suitable applicant would possess either a strong grounding in distributed
systems or a strong mathematics and computing background.

For more information contact:

Howard Bowman (hb5@ukc.ac.uk)
or 
John Derrick (jd1@ukc.ac.uk)


                 Visual Programming for Parallel Programs
                 ========================================

Visual programming aims to ease the task of programming and make it more
accessible to a larger population. This is done by specifying the
program via visual expressions (such as graphics, drawings, animation or
icons).

The design of such a language has been defined for a visual editor of 
parallel programming languages. The editor provides visual expressions
for a number of parallel processing methodologies but aims to be
language independent, as a programming editor should be.

The student will be expected to implement this editor and provide
research input to the project. Proposed stages of the project:

	i)   Implementation of the prototype visual editor;

	ii)  Research into methods of mapping visual expressions onto
	     textual language constructs;

	iii) Design and implementation of the general user interface;

	iv)  Constructive comment of the use of the editor, as a 
	     serious tool for parallel processing;

	v)   Extensions to the editor to enable use of programming 
	     distributed systems.

The student is expected to acquire skills in user interface design,
graphics programming and general parallel programming.

For further information contact:

Andrew Smith (abs@ukc.ac.uk)


                  A Multiple Precision Arithmetic Package
                  =======================================

Many numerical problems faced by physicists and engineers require the use
of arithmetic whose precision is far greater than that provided by
standard floating-point hardware. A number of packages exist but they are
mostly written in Fortran 77 and are extremely unwieldy to use.

The object of this project would be to write a multiple precision
arithmetic package using Fortran 90, the new Fortran, standardized just
three years ago. This language is far removed from its predecessors
and provides a number of features, like the definition of operators, which
would make such a package very easy to use. Not only would the basic
arithmetic operations be implemented but also a wide variety of common
mathematical functions (for example, sin, log, exp, etc.).

One feature of the project would be extensive testing of the basic operations.
This would include comparisons of results between those obtained by the
package and those generated by computational algebra packages, which tend to
be extremely slow for these operations. (The Computing Service provides access
to both Maple and Mathematica.)

It would be expected that the codes generated during the project would be
suitable for publication in a journal like ACM's Transactions on Mathematical
Software.

For further information contact:

Tim Hopkins (trh@ukc.ac.uk)


         Functional Programming and Scientific Visualization
         ===================================================

The aim of the project is to explore the application of functional
programming languages and techniques to the area of scientific
visualization.	In particular, the use of laziness both in the small - for
fundamental algorithms - and in the large - for module inter-connection -
is seen as an exciting avenue for research.

The benefits of the functional approach are:

+ An improvement in efficiency may be obtained from exploiting
  the lazy nature of functional languages.  Only those items of data
  required for a particular procedure will be evaluated.
+ An improvement in memory usage may be obtained and data may be
  shared safely between system components.
+ Design of visualization components (or modules) will be facilitated.
  Modern functional languages provide powerful mechanisms for the
  description of generic components, without the need for run-time
  type checking.
+ Design of visualization applications will be facilitated
  by the secure module systems and higher order nature of functional
  languages.
+ Systems will be smaller due to the concise, semantically
  clear nature of functional programs and, in particular, because memory
  allocation and de-allocation will not be a concern of the programmer.
+ Functional languages are mathematically elegant and amenable to
  formal transformation and verification.  A functional visualization
  system will be robust.  Properties of critical numerical components will
  be amenable to formal verification.

For more information contact

Steve Hill (sah@ukc.ac.uk)


                  Parallel Functional Programming
                  ===============================


The aim of this project is to build an implementation of a
explicitly parallel functional programming language on our
Meiko Computing Surface - a transputer-based MIMD computer with
over 120 processors.  The work would be based on the Gofer language
(similar to Miranda and Haskell).  Gofer is written in C, so a
good degree of proficiency in this language would be required.
The project would be staged as follows, although it may not be
possible to complete all stages in the time scale of the project.

+ Port the standard version of Gofer to the Meiko machine.
+ Research and experimentation into parallel functional
  languages, and the facilities provided for parallel
  programming and process and channel creation on the Meiko.
+ Implement parallel facilities on a single processor.
+ Implement multi-processor parallelism.

Applicants should have some knowledge of parallel processing, 
functional programming and general software enginnering.

For more information contact:

Steve Hill (sah@ukc.ac.uk)

		Plain text to marked-up text
		============================

With the wide availability of scanners and optical character recognition
software, it is easy to obtain documents in plain text form.  Such
documents have lost all their structure: there are no sections,
paragraphs or emphasis for example.  The aim of the project would be to
investigate ways in which the structure of the original document can be
inferred from the layout and content of a plain text document.  The
structure of this derived document should be expressed using a mark up
language such as SGML, HTML or LaTeX.  To limit the scope of the project
it is suggested that a converter be written for a restricted set of input
documents such as bibliographic references, rather than English text.

For more information please contact:

David Morse (drm@ukc.ac.uk)
or
Judith Wusteman (jcw@ukc.ac.uk)

	Support for cooperative authorship of documents
	===============================================

It is not unusual for two or more people at different institutions to be
working on a document.  While documents written in LaTeX or groff can be
exchanged easily by electronic mail, the annotations and corrections
which the other authors make to versions of the document cannot.  The aim
of the project would be to investigate ways in which documents written
using document preparation tools such as LaTeX or groff can be displayed
with annotations which another author has made to the document.  

Conventions would have to be developed so that annotations such as
deletions, insertions and marginal notes could be indicated in the
electronic version of the document.  This could take the form of extra
macros if LaTeX or groff were used as the document preparation system for
example.  The marked up document could then be converted to a form in
which it could be viewed using the Adobe Acrobat viewer.

For more information, please contact:

Judith Wusteman (jcw@ukc.ac.uk)
or
David Morse (drm@ukc.ac.uk)

		    A hand-held navigator
		    =====================

This project would investigate the feasibility and utility of an
electronic map to aid the pedestrian navigating around a strange building
or city.  The map would be displayed on a hand-held computer such as an
Apple Newton and should, as well as displaying the map, be capable of
giving directions such as `Go up the stairs and turn left at the top', or
`Go down the next street on the right'.  As well as having a complete map
of the locality, the navigator should be able to reason about the map and
be able to work out, for example, the shortest route between any two
points.  In addition, the navigator would need to know its current
location: this could either be enquired of the user, or preferably,
obtained automatically from the environment via sensors used indoors in
`Active Badge' technology or using Global Positioning Systems in the open.

For more information please contact:

David Morse (drm@ukc.ac.uk)