You Need a Psychologist to Teach HCI Correctly to a Computer Scientist

For the first time since its inauguration, IFIP's conference on human computer interaction, INTERACT, made its way to the southern hemisphere. INTERACT'97 was held in Sydney, Australia, in July, and incorporated two regional conferences, OZCHI97 and APCHI97. The theme of the conference was "Discovering New Worlds of HCI" which drew participants from all over the globe and from a wide range of disciplines.

It was this range of disciplines which set the scene for a lively discussion in the opening panel session where the proposition was put forward that a psychologist is needed to teach HCI correctly to computer scientists.

If HCI is to be successfully adopted and practised in industry, software developers will require some form of education or training in this area. The question remains as to who should be providing this teaching -- a psychologist, or a computer scientist? The panel and audience grappled with a number of questions during the course of the session. As well as the question of exactly who should be responsible for the teaching of material, other related issues were discussed including who should be responsible for the content, what material should be presented, and how it should be taught.

In attempting to address the main question, it was not surprising that the related question of what needs to be taught to these individuals was raised a number of times. While it is not possible to teach computer science students everything they will need to know about HCI, it was recognised that a key objective is to increase the awareness of students of their own limitations and the boundaries of their discipline so that they learn to ask important questions and raise relevant issues during the course of their practice. Ideally, the student should be taught to draw from the knowledge provided by external disciplines and be able to apply it to the problem at hand. The material taught needs to be well-focussed and provide students with a strong sense of what HCI is about and what it tries to achieve. It needs to be directly relevant to the capabilities and needs of the students to whom it is presented, and able to be used in industry both currently and in the long term.

Computer science students by their nature were recognised as having a rigid approach in their thinking and reasoning. This can give rise to incompatibilities between the more qualitative nature of psychological knowledge and the more quantitative heuristics needed by software developers. While computer scientists may be able to communicate to their students at this level, there is a danger that they may misapply psychological knowledge or oversimplify human behaviour. An example was given where a computer scientist had taught a class that graphical user interfaces were bad when they had more than seven items which was thought to be in conflict with Miller's recommendation for short-term memory.

It was pointed out by both sides that no single person is likely to be able to give a comprehensive picture of HCI as a whole. Computer scientists are likely to have a poor understanding of areas in the social sciences. Psychologists are likely to have a poor understanding of the software engineering context required to build systems. Even multi-disciplinary academics and practitioners can have their own specific prejudices due to their limited perspectives on HCI. So where does the solution lie? Two general approaches were offered. The first makes use of a team of individual specialists to provide a broad scope. The second recognises HCI as a discipline in its own right which needs to be taught by those who understand HCI as a whole, rather than just specific components. However, neither of these approaches were seen as being particularly easy to implement.

Finally, the question of how the material should be taught was also posed. The notion of trying to integrate the teaching and assessment of HCI components throughout the entire curriculum rather than just in isolated units was supported. For example, a number of marks could be specifically allocated towards usability in programming assignments. In addition, there was a recognised need to place students in large and realistic projects, possibly as a result of partnerships between industry and universities.

There appears to be significant parallels between HCI education and the HCI discipline as a whole. Firstly, as HCI struggles to define itself, HCI education needs to focus its scope and objectives. Secondly, as industry and academia look to match results and needs, students need to distinguish between what HCI can do, and what it does in practice. Thirdly, as each of the HCI component disciplines search for common ground, students need to be taught to recognise the boundaries and be able to apply the relevant information to the problem at hand. With such high goals, perhaps a more pertinent question is, "Who is the most appropriate person to teach HCI to the teachers?"

Panel Organiser: Marilyn Mantei, Canada
Chair: Penny Collings, Australia
Panelists: Tom Hewett, USA; Andrew Monk, UK; Lorraine Johnston, Australia; Mark Apperley, New Zealand

Paul Chesson is completing a Ph.D. at the University of Melbourne, Australia, in the area of formal methods in HCI. His research involves the development of a dialogue specification language for interactive systems, and a supporting method and tools for its use. He holds a B.E. from Melbourne in software engineering. Email: chesson@cs.mu.oz.au

Issue

Article

Vol.30 No.1, January 1998

Article

Issue