Prof. Andrew Basden
A.Basden@salford.ac.uk
Andrew Basden,
Professor of Human Factors and Philosophy in Information Systems,
I.S. Institute,
University of Salford,
Salford,
Manchester.
M5 4WT.
Tel +44 161 295 2913
(Please leave a message
when I'm not there.)
Fax +44 161 745 8169
Room F18,
Newton Building.
ext 52913.
Full Curriculum Vitae
|
After obtaining a First Class Honours Degree in Electronics (Science) at the University of Southampton in 1969, Andrew Basden entered research in the field of Computer Aided Design. For his PhD Thesis he pioneered a topological method for laying out electronic circuit boards that still has relevance today, and discovered the creative joys of writing software. He spent the next twelve years outside academia in the business uses of computers - starting with data analysis, progressing through medical records (and the problems of error-ridden data) into expert systems (knowledge based systems; KBS) in the chemical industry and the surveying profession. The experiences during this period impressed upon him that technology is people, not just computers, and he started to devise approaches that took into account the real-world needs and benefits. Returning to academic life in 1987, as lecturer in the Information Technology Institute at the University of Salford, he started to research the foundations of this experience, devise models and methodologies, and seek philosophical underpinnings. He has taught a wide variety of topics, and has contributed via his research to the following areas, spanning both 'soft' and 'hard':
|
This area of research covers research and development of technologies. Andrew Basden's main interests are in technologies that represent knowledge and provide the user interface.
During the first part of the 1980s Andrew Basden worked in ICI plc on knowledge based systems. In 1986 he joined the University of Salford, as chief Knowledge Engineer for the highly acclaimed ELSIE project which became one of the most successful KBS of its time. The ELSIE project built upon and refined his ICI experience, and led to a number of funded projects to develop the Client Centred Approach to building information systems and knowledge acquisition methodology.
But Andrew was also concerned that the shape of KBS technology should be made more 'appropriate' to the needs of the real world and real-life knowledge representation. This concern led him to a number of related areas of research:
- Theory of appropriateness.
What does it mean for a formalism to be appropriate? (Part answer: the symbol level facilities should match the knowledge level meaning.) And why is it necessary? (Part answer: inappropriate knowledge representation formalisms make these processes clumsy, inefficient and error-prone.)
- Integration of KR formalisms
. There are several 'aspects of knowledge' in real-world applications - conceptual, quantitative, spatial knowledge and activity - that are irreducibly distinct and difficult to integrate. The theory of appropriateness gives us a basis for doing so that can guide design of software. Spatial knowledge, in particular, has not been properly integrated, so, during the 1990s, Andrew was involved in three funded projects to investigate how this might be achieved. (These four 'aspects of knowledge' are now seen to be a subset of the philosophical aspects below.)
- IRKit - Kit for Integrated Representation of Knowledge.
This is on-going, unfunded research to devise a suite of modules designed according to the principles of appropriateness and of levels, to see whether the approach has merit. Assembly Language is employed, not so much for speed, as to illustrate the principle that appropriate symbol level facilities of the various aspects should be implemented at the bit level. IRKit has been used in the construction of ...
- ... the Istar KBS toolkit.
Istar is working, robust software designed to show the practicality of a number of research principles mentioned here, including:
Istar has proved remarkably successful in allowing the creation of knowledge bases in ill-structured knowledge, such as for contract authoring, evaluation of sustainability, etc. and has even been used as a cognitive mapping tool.
Istar has been expanded so that it can be run as a server on the Internet, making expert knowledge available worldwide via simple web browsers. It has been designed so that it takes all the load of multi-threading, sockets, connections, dynamic HTML page construction, timeouts, etc. so that knowledge bases need no modification to run in server mode. This technical possibility has opened up new questions about cross-cultural distribution of knowledge, and research in both this and technical features is currently underway.
Several demonstration knowledge bases (including contract authoring) are available on the Istar Knowledge Server web page.
Some of the technical issues in adapting a KBS inference engine to use on the web are discussed in:
Basden A, (2000), "Some technical and non-technical issues in implementing a knowledge server", Software - Practice and Experience 30:1127-1164.
Said Donald Norman (1990):
"The real problem with the interface is that it is an interface. Interfaces get in the way. I
don't want to focus my energies on an interface. I want to focus on the job."
It became clear that when building complex knowledge bases in ill-structured domains the process is more akin to creative design than to assembly of knowledge-pieces. The user interface must impose minimal cognitive effort and not interrupt the continuous flow of thinking that is taking place. This is important in software tools that aid creative thinking and design and for immersive software like computer games and virtual environments. The traditional point-and-click 'graphical' user interface is too 'distal' for these.
Using ideas from the philosopher Michael Polanyi, Andrew Basden has devised the notion of the Proximal User Interface (PUI), worked out a set of principles for design and evaluation of PUI, and implemented and tested these in two software packages (Istar knowledge based system toolkit, and Annotator , which allows annotation of images). What is new about PUI is not so much the technology, as that it is a new way of thinking about user interface as such: no longer seen as an intermediary as a 'glue' that fits the artifact comfortably to the user's mind.
How do we encapsulate knowledge in the computer? Normally we 'engineer' or 'build' it, or 'write' it. Why not draw it, instead? Just as we do on the proverbial back of the envelope, we doodle and scribble, and by doing so set down meaning.
Traditional drawing packages might allow us to produce good pictures - but they do not capture the meaning of what we draw. Visual programming packages capture some of the meaning - but only entities, attributes and relationships and seldom such things as complex spatial knowledge as found in contour maps, surface coverage maps, etc., and in any case they are rather clumsy, 'distal', in use. What we need is:
- something highly proximal so that drawing becomes part of an uninterrupted thinking process,
- something that can capture the diverse types of meaning that we wish to set down - not just entities, attributes and relationships.
Andrew Basden is supervising research (by Kamaran Fathulla) into a new approach to 'thinking by drawing' - including both establishing the underlying theory and also implementing it in software. This research goes in the reverse direction to that normally encountered that is directed at how to create displays for existing data or knowledge or to undertake spatial logic. The theory involves both that of Proximal User Interfaces, appropriateness and philosophy.
So far, this has been investigated in two contexts, the creation of knowledge bases in ill-structured domains (Istar) and the annotation of historical images (Annotator). In both, the user draws what they mean and immediately what is drawn can be searched or executed.
This area of research concerns how to use the given technologies (whatever shape they come in) to produce artifacts for use by people in real life situations. Andrew's research is oriented towards building knowledge-based artifacts.
Having had to build real-life KBSs over the years, Andrew has been forced to find ways enhancing the quality of knowledge acquisition. While most researchers sought ever more sophisticated acquisition techniques, he directed his attention instead to the quality of what was being acquired. In this way, he devised a methodology that helps the knowledge engineer avoid the well-known problem of brittleness, and also to turn disagreement between experts from the serious problem it is traditionally taken to be into a source of new high quality knowledge. This work, started in the 1980s but still quite novel and continuing at a low level, is described in:
Attarwala FT, Basden A, (1985), "A methodology for constructing Expert Systems", R&D Management, v.15, n.2, pp.141-149.
However, there is a further problem: KBSs are often far too narrow in the knowledge they contain. Andrew supervised, and then worked with, research by Mike Winfield (University of Central England), to produce the MAKE (Multi-Aspectual Knowledge Elicitation) method that has proved successful in stimulating experts to consider all the aspects of their work and also in explicating tacit knowledge. Its strength comes from its philosophical base. This work is reported in:
Winfield M J, Basden A, Cresswell I, (1996), "Knowledge elicitation using a multi-modal approach". World Futures 47:93-101.
and briefly described on Andrew's website in MAKE.html.
It became clear to Andrew, during his work in the 1980s and early 1990s, that knowledge acquisition and appropriate knowledge representation formalisms were not enough on their own to ensure success of a KBS. Often superb KBSs fail because of 'political' factors within organizations or flawed development methodology. To this end, Andrew was research manager in the £1m EDESIRL project that produced the Client Centred Approach (CCA) and Client Centred Methodology (CCM) for constructing information systems in which knowledge is important.
CCM combines a linear with an iterative structure to be both responsive to changing client needs and situations and controllable within set timescales and resources. It manages this because CCA pays attention to the multi-stakeholder human side of real-life development projects, and to the need to aim for real benefits rather than just delivery of technology. It contains a strong emphasis on ethics. These are described in
Basden A, Watson I D, Brandon P S, (1995), Client-Centred: An Approach to Knowledge Based Systems, CLRC: Rutherford Appleton Laboratory, U.K. ISBN 0 9023 7635 7.
Watson ID, Basden A, Brandon P, (1992) "The client-centred approach: expert system development", Expert Systems, v.9, n.4, pp.181-188.
Watson ID, Basden A, Brandon P, (1992) "The client-centred approach: expert system maintenance", Expert Systems, v.9, n.4, pp.189-196.
Of course, an important element of such a methodology is to take account of real-world usage, both desired and actual.
This research area is concerned with the I.T. artifact being used within human situations. It is closely linked to parts of development methodology, but is a distinct issue since we are dealing with users rather than developers as the primary human beings involved.
There is a difference between ease of use and usefulness; greater ease of use can lead to time-wasting, shoddy thinking, etc. What is usefulness, and how does it relate to technical features?
A joint EPSRC-funded project with the University of Newcastle Psychology Department studied the successful KBS, ELSIE, in use among quantity surveyors. A particularly significant research outcome, described in
Basden A, (1994), "Three Levels of Benefit in Expert Systems", Expert Systems, v.11, n.2, pp.99-107,
was the development of a three-layer model of usage and benefits, by which the link between technical features and impact the artifact might have when in use can be understood, and hence predicted, evaluated and designed for. It has been applied, for example, to virtual environments technology.
With $1,500bn spent every year on information and communication technology, is the world receiving $1,500bn worth of benefit from that spend? Many estimates suggest that the failure rate of information systems is in the order of 70%.
It is not enough to understand how impacts of technology use come about; some impacts are positive, leading to success of an information system, while others are negative. What might be benefits to one person is detriment to another. Benefits might accrue in the short term and immediate context, but there might be indirect, unintended and unanticipated detriment in the longer term. How do we address these issues?
Andrew is working on a new approach that starts from a philosophical foundation that suggest new frameworks for understanding I.T. usage, success and failure. It is briefly outlined in the section in 'Beyond Emancipation' on Distinguishing Success from Failure.
This area of research is concerned with how we see information systems, and requires philosophy:
Starting from knowledge based systems, and the KBS community's attempts to come to terms with knowledge, Dr. Basden has extended Newell's (1982) concept of the knowledge level to a full suite of Levels of Meaning:
- Tacit (Context) Level - cultural assumptions
- Knowledge Level - the application content
- Symbol Level - the symbol structures
- Bit Level - signals and information resources
- Component Level - hardware
- Materials Level - materials technology
which are irreducible to each other such that each covers a range of topics relevant to information systems and their use. For example, for a system to be successful it must work well at all levels. This suite parallels those in linguistics. It has been particularly useful in teaching (allowing phenomena at different levels to be kept distinct and ensuring that all levels are taught), and has informed research such as the IRKit architecture.
How do we bring all the above interests together in a coherent way that underpins them all? According to Hendrik Hart [1984] philosophy is the integrative discipline. To this end, Andrew Basden is exploring the pluralist philosophy of Herman Dooyeweerd. The way it can underpin all areas is briefly discussed in 'Christian Philosophy and Information Systems. What seems to be significant about this philosophy is that it is able to underpin all the main areas of concern: technology, methodology, usage and perspectives. Very few other philosophies seem able to do so.
A major element in Dooyeweerd's philosophy is his pluralistic ontology of aspects that allows us to avoid both the monistic-reductionist and the dualistic approaches that others give us, that is useful in both methodology and usage. It also parallels the levels and the aspects of knowledge, and is the basis of MAKE. The aspects can provide us with an evaluation tool that can stimulate discussion:
Andrew Basden is a founder member of the Centre for Philosophy, Technology and Social Systems, jointly with the Free University of Amsterdam and the University of Luleā, Sweden, and is 'gatekeeper' for The Dooyeweerd Pages.
It seems this philosophy can bring together the three main stances on information systems: 'hard', 'soft' and 'critical'. Tentative discussion of this is contained in 'Beyond Emancipation', which asks the question: if 'soft' arose dialectically out of the problems of 'hard', and 'critical' out of 'soft', what next? How can we avoid mere wandering around the dialectic landscape for ever? It may be that Dooyeweerd's ideas can help.
'Interdisciplinary' is a much-used word today - but what does it mean? It has first signified a desire to escape the narrow confines of single, isolated disciplines, but in most cases it has then meant merely pushing two or more disciplines together, hoping for some synergy. Both information technology usage and environmental sustainability and planning are interdisciplinary in nature, so there is a need for a more principled approach to them.
Andrew has sought for a sounder basis for interdisciplinary thinking and working. Dooyeweerdian philosophy provides one such basis, postulating a clear yet theorectically robust view of what a discipline is and what is the relatiohship between them. This work is in the early days, but it is hoped that a firm theoretical foundation can be found that is strong enough to support a wide range of interdisciplinary activity.
Page completely rewritten 17 December 2001 by Andrew Basden. 8 June 2005 title.
The older, 1997, version is still available.