CHI 97 Electronic Publications: Plenary and Invited Talks
A Typology for Educational Interfaces
Tim O'Shea
The Open University
Milton Keynes
MK7 6AA, United Kingdom
T.M.M.OShea@open.ac.uk
ABSTRACT
Interfaces intended to support learning should be considered with respect to a typology based on student audience, constructive functionality, navigation support, cognitive cost and added learning value. Analysed like this, the quality of interfaces used by students has noticeably improved over the past 10 years, in dramatic contrast to the much slower change in pedagogic value of educational software. The potential for the use of computers in support of interaction between learners, their peers and remote information sources has revealed important weaknesses inherent in current approaches to navigation support. Key problems include scaleability, accessing peer learners and the shape and size of information spaces.
Keywords
Educational interface, navigational support, memory prosthesis, scaleability
© 1997 Copyright on this material is held by the authors.
INTERFACE DIMENSIONS
I approach user interface design from the point of view of improving the quality of learning support. Many interfaces are designed for use within a very specific educational context and may be designed to support the acquisition of a particular skill or access to a limited information source. However the general variation in learning styles, learning chronology, motivation and personal history of information technology use is so great that in order to assess the pedagogic utility of an interface it is necessary to situate particular types of interface design in relation to each other on a number of orthogonal dimensions. This paper focuses on 5 dimensions that can be used to analyse the relative strengths and weaknesses of extant educational interfaces and to focus on the key design issues for the next decade.
STUDENT AUDIENCE
Many designers make the vacuous claim that their interfaces can be used by anybody. Given the potential reuse of educational software in different settings this claim has to be approached with complete scepticism. Most interfaces are not usable by users below a certain age or reading level. Many interfaces are not usable by learners that have visual or hearing impairments. At the Open University we take particular pride in the way we extend access to higher education by designing interfaces for students with special needs and the 6,000 such students in our community are a critical forcing function for us. The Virtual Microscope [1] was designed for students who would have difficulty with our standard physical microscope but has proved valuable for a much wider range of students. Students vary according to their gender, social background and academic affiliation and a style of computer generated explanation (or joke or cartoon) that might work well with a young female working class engineering student could also be irritating, opaque or unintelligible to an older male middle class classics student. A critical aspect of student variation is their comfort with and previous history of information technology use and their prior knowledge of the topic they are using the interface to learn. A part-time university student may study for 6 to 8 years and the point at which they use a particular interface will mean that some will be nervously learning to use it for the first time and others will be experienced (or over-experienced) with the relevant interface facilities or features.
CONSTRUCTIVE FUNCTIONALITY
Educational technologists need to be clear exactly what students can be expected to construct with a given interface. The additional support given to learners will depend on whether they are constructing database queries, messages, search specifications, parameters for a simulation, models, their own database, pictures, tunes, animations, movies or programs. Often interfaces are so heavily decorated with attractive icing that it becomes difficult for the student to focus on what is being built when the interface is being used. This becomes especially critical when a single interface supports a multiplicity of different types of construction activities. Interfaces should be designed to make it easy to retrieve a compete history of use. The student should be able to easily relate how they built what they built to subsequent system behaviour.
NAVIGATION SUPPORT
Learners get lost in many of the different types of information space we give them access to. They often do not understand the shape of the space which maybe a 2 dimensional plane, a tree, a network, a lattice or some odd or subtle topology. Learners may want to get back to a point in the information space they have previously visited or locate some other learner who is simultaneously active in the space. The navigation support tools currently available make strong implicit assumptions that learners understand difficult structures such as hierarchically organised trees or upper semi-lattices. Also they do not support some types of navigation that are vital for effective learning. Sometimes browsing is fine but often a learner will need to plan and organise their movement in an information space over days and weeks in support of a research project. They may wish to 'time travel' back to a place and learning activity carried out with other co-learners in a previous semester. They may need to travel in parallel in different information spaces, adding to a database structured in one way recording runs of a simulation structured in another. Smith's Shared-ARK[2] system has a navigation tool modelled on radar which supports learners working together on an apparently infinite plane.
COGNITIVE COST
A student can use a number of different interfaces in a semester. A student who is doing nearly all their work electronically may use more than a hundred. They will have to deal with interfaces that are designed with underpinning from different physical, mechanical, locational metaphors. Some are essentially self-referential and can be thought of as computer-computer metaphors [3]. They are not always accessible to the learner. In some interfaces the metaphors are badly mixed and in many cases they break easily. Shortcuts conventions and syntax varies and learners have to deal with icons and acronyms whose cuteness is more than matched by their opacity. These variations cost the learner and a particular burden comes from interfaces that are almost but not quite identical. Learners need consistency and reliability and an important new cognitive demand on them stems from the variation in response time of search engines deployed on remote machines. Learners also need memory prostheses so that they can reduce cognitive costs by avoiding the burden of remembering the fine detail associated with individual successful learning interactions.
ADDED LEARNING VALUE
We need to take a wide view of education and having checked the value of an interface in support of the learning goal for which it is being deployed we must also consider what added or incidental learning value might be associated with it. Many interfaces support visualisation of abstract ideas and the use of multiple representations and this style of working can be generalised into personal higher order learning skills. Many interfaces are augmented by redundant sound, colour and animation [3] and this can have a very beneficial effect on the pacing of work and on motivation. Interfaces that support co-operative working are a great benefit for physically isolated students and also help learners develop their interpersonal and communication skills. The navigation tools and memory prostheses that work well help the students explore more widely and build learner autonomy.
THE LAST TEN YEARS
There has been dramatic progress in constructive functionality. Student inputs that would have been expressed as typed sentences of keywords and processed by clumsy parsers are now conveyed by economical and elegant mouse gestures. Learners can now deploy speech, graphics, sound, animation and video with much more ease and the range of potential students is being steadily extended up and down the age range and, critically, to a wider range of students with special needs. In addition to this modest improvement in educational reach to student audiences, there have also been enhancements in the general added learning value associated with some types of interface, especially those that support peer learning and co-operative working such as those associated with the Virtual Summer school [4]. However the cognitive costs to the learner are rising and one good measure of this is the increasing amount of material and activity that is focused on supporting learning about interfaces and providing remedial help during electronically mediated instruction.
FIVE ISSUES FOR THE NEXT TEN YEARS
The most pressing problem is the need for improved navigational support for learners. Encouraging students to access the Internet, explore very large databases or work in powerful simulation environments without the equivalent of maps and compasses is counterproductive. Interface designers and educational technologists need to be more disciplined in their choice of topologies for information spaces and specify usable navigation tools. Second, we must take a holistic approach to the learner's experience and find ways of reducing cognitive costs by selecting and remaining consistent with a small number of general metaphors and shortcut conventions in interface design. Third, we need to develop effective memory prostheses that support the learner in recalling the fine detail of an increasing volume of electronic interaction. Fourth, we need to ensure that approaches that work for small numbers of learners are scaleable and work is settings such as the Open University which has courses in which 5,000 students will be using the same simulation or conference for the same learning purpose in the same week. Future added constructive functionality will be mostly driven by wider business processes but it will be necessary for educators to continue to search for opportunities to widen student audiences and to increase added learning value. A key focus for this final issue are students with special needs who may find new facilities potentially reduce value. A compelling example is the initial impact of windowing systems on blind students[5]. The challenge for the future is to provide students with useable interfaces that support higher quality learning while extending access to the new technologies.
REFERENCES
1. Bennett S., Metcalf, J., Ross, S., Scanlon, E.,Thomas, J., Williams, D. Opening up science: the teaching of science at the Open University, United Kingdom In Sewart, D. (ed) Proceedings of 17th World Conference of the International Council for Distance Education, pp 9-17, 1995, June Birmingham
2. Smith, R., O'Shea, T., C. O'Malley, C., Scanlon E., and Taylor, J. 'Preliminary Studies with a distributed multi-media problem-solving environment' in 'Studies in Computer Supported Co-operative Work' (Eds. J.M. Bowers and S.D. Benford) North-Holland 1990.
3. O'Shea, T., O'Malley, C. and Scanlon, E. 'Magnets, Martians and Microworlds: learning with and learning by OOPS'. Journal of Artificial Intelligence in Education, 1(3), pp11-26, 1990.
4. Eisenstadt, M., Brayshaw, M., Hasemer, T. and Issroff, K. Teaching, Learning and Collaborating at an Open University Virtual Summer School. In A. Dix and R. Beale (Eds.) Remote Cooperation: CSCW Issues for Mobile and Teleworkers. London: Springer, 1996
5. Edwards, A., and O'Shea, T. 'Making graphics based programming systems usable by blind people' in Interactive Learning International, 1985.
BIOGRAPHY
Tim O'Shea is Professor of Information Technology and Education and a Pro-Vice-Chancellor at the Open University. He has a B.Sc. from Sussex, Ph.D, from Leeds and has worked at Texas (Austin) and Edinburgh Universities. He has held vsiting appointments at Xerox PARC and Berkeley. He has published 9 books, more than 100 articles and a 6 part BBC TV series on computer assisted learning, artificial intelligence and distance education.
CHI 97 Electronic Publications: Plenary and Invited Talks
