Issues in Computer-Supported Collaborative Learning
Chapter to appear in Proceedings of NATO Advanced Workshop on Computer-Supported Collaborative Learning (Claire O’Malley, Editor) held in Maratea, Italy, Sept. 1989.
Liam J. Bannon
Computing Centre
University College Dublin
Ireland
bannon@irlearn.ucd.ie
"...various combinations of medium/context relations (which are themselves mutually constituting) reorganize the conditions for people to coordinate with each other."
Introduction
This paper presents some personal observations and critical reflections on the emerging field of Computer Supported Collaborative Learning (CSCL). The perspective adopted here is not based on a deep immersion in this new field, nor does it come from some omniscient position outside. Rather, I would regard myself as a fellow traveller, in that the concerns of a number of the people who identify with this area overlap with my own. For example, we share a common concern with how learning occurs, and can be supported through various media, especially computers. More specifically, we also share an interest in collaborative processes as distinct from individual ones, whether it be in the context of learning or more general work settings. The Chapter is organized as follows: In Section 1 I discuss some issues concerning the nature of the research conducted under the rubric of CSCL, with a view to determining what are some of the crucial issues in the field. In Section 2 I describe some specific examples and uses of technology which serve to highlight certain issues in the field. Finally I give some brief comments on how these observations may be instructive in envisioning the design of computer systems to support collaborative learning.
1. The Nature of CSCL
What is CSCL? There are at least two quite different approaches to answering this question, and we will look at them in turn. One perspective views the CSCL term as simply an "umbrella term" which serves a useful function by bringing together under its umbrella, in meetings and workshops, a variety of researchers with different backgrounds and techniques, where they can discuss their work. This allows for the cross-fertilization of ideas, for the fostering of multi-disciplinary perspectives on related problems. Indeed, such a stage is often a prerequisite to the further development of many new research areas, as I have argued elsewhere, in discussions about the development of the related area - Computer Supported Cooperative Work (CSCW) (Bannon et al.,1988, Bannon & Schmidt, 1991). In this approach, CSCL is simply whatever people who come to these meetings say it is. In a sense I think we are at this early stage in CSCL, as judged by the variety of approaches and views that we find in the current volume.
An alternative approach to understanding CSCL is to try and ascertain exactly what are the unique problems and concerns that might make CSCL a separate problem area for researchers. By having some shared concepts and some agreement about what the object of study is, the field could develop more coherently, rather than simply exist as a marketplace for quite different ideas. Taking this view, what do we understand by CSCL? Since it is obviously not a unitary concept, perhaps one way of approaching an understanding of the term is by a compositional approach. We build up the meaning of the term from its components. Thus we can ask, in this field, what do people mean by Learning (L), by Collaborative Learning (CL), by Support for Collaborative Learning (SCL), and by Computer Support for Collaborative Learning (CSCL)? The issue here is not to impose an exclusive interpretation on what is or is not meant by the term CSCL , but to bring our attention to the object of interest (or do we mean subject?) - namely learning, specifically collaborative learning, and how it might be supported by the computer.
1.1 Learning vs. Schooling
To keep these notes short I will not indulge in a detailed step by step exegesis of CSCL as laid out above, although I do believe that an analysis of how the different authors view the very nature of "learning" itself would be instructive. I would like to make a few points though concerning our object of analysis in this field. Much of the interest in learning and computer support of learning is focussed on the schools, not surprisingly. However, given the debate about how effective, or rather ineffective such establishments are in producing "learning", the temptation to equate learning with schooling should be resisted. While the emphasis of much current CSCL work is in classroom interaction, we should acknowledge that much learning goes on outside of the formal classroom, and that indeed it is precisely in such contexts, outside the classroom, that CSCL could make a significant impact. This debate about the misleading role of educational institutions in "educating" people has re-focused attention on the importance of everyday social practices of people at work and play that afford opportunities for learning (Lave, 1988). This is part of a broader critique of our understanding of how learning occurs, that discredits older instructional paradigms (e.g. Gagne, 1968) and argues for the social "situatedness" of human learning and action (Suchman, 1987, Winograd and Flores, 1986). This alternative approach questions the separation of "formal" and informal" education, and argues for the importance of "apprenticeship" in learning (Lave, 1977). It includes "outside-class" activities as an integral part of the social context in which learning occurs (Cole & Griffin, 1987).
1.2 Learning as a Collaborative Activity
The focus of much learning research has been on individual learning, coming from a behaviouristic perspective. The role of others in this learning process was often regarded as ancillary at best. More recently, such approaches to how learning occurs are being challenged. Of particular note is the inventive empirical work of researchers who base their theoretical framework on such people as the Russian sociocultural psychologist Vygotsky (1978) and others, for example, Leontiev (1978). A key concept in this work is Vygotsky's idea of the "zone of proximal development" (ZPD) as the site where learning occurs. The zone is formally defined as: "the distance between the actual developmental level as determined by independent problem solving and the level of potential development as determined through problem solving under adult guidance or in collaboration with more capable peers." (Vygotsky, 1978, pg 86). This concept has been at the heart of the educational research of several groups that is very relevant in the current context, as they all stress the importance of studying learning as a collaborative process, and they have also used the computer as a medium in which to provide new contexts in which this collaborative learning might take place (e.g. Newman, Griffin & Cole, 1989). Greater emphasis on field research and actual observation of joint problem solving activity is evident. In this volume, we can see several papers, besides this one, that are influenced by this framework (Crook, Newman).
It is difficult to view the studies on collaborative learning as a whole, however, as the studies themselves often differ radically in both the theoretical perspective adopted and in the research paradigm used. Let us just note here some of the differences in the research on "collaborative learning" . Who or what is collaborating with whom, and under what conditions? While most papers do not address this issue specifically, the range of meanings appears quite broad, in that for some it consists of a single individual collaborating, not with another person at all, but with a computer system (e.g. Dillenbourg, this volume)! For others it is the study of 2 randomly-chosen subjects collaborating on solving an artificial task, chosen by the experimenter, on a computer. In other cases, those studied are class mates, who at least have some shared experiences together. Other research focuses on how large ensembles of individuals learn to coordinate their actions through the computer over a period of months. It seems patently obvious that the kinds of cooperation evinced, with or without the computer may be radically different in these different situations, yet to my knowledge we do not have any agreed framework for comparing and contrasting these studies on collaborative learning. Just as an indication, some features that any such framework would need to cover include:
The nature of the collaborative task : puzzle solving, editing a newsletter
The nature of the collaborators: peer, teacher-student, student - computer
The number of collaborators: 2, 3, 100 ( e.g. a shared hypertext system with entries by many people)
The previous relationship between collaborators: how many shared experiences
The motivation for collaboration: intrinsic interest, experimenter defined task, money
The setting of collaboration: classroom, workplace, home
The conditions of collaboration: physically co-present, computer-mediated
The time-period of collaboration: minutes, hours, days, weeks, years
Etc.
The issue here is not to argue for any particular view of what the prototypical CSCL scenario might or should be, but to be aware of the quite vast differences in meanings among researchers, even when they utilize similar vocabularies about collaborative learning. This is important because what counts as collaboration or cooperation between people differs widely in these different prototypical setups. For example, does it imply shared goals? Is collaboration or cooperation simply opposed to competition, or does it imply more ?
1.3 The Role of the Computer
Looking at cognitive change and learning processes, Kurland & Kurland (1987) refer to Collins (1984) in discussing ways the computer can be used to support learning. It can allow one to simulate situations that would be impossible in the real world, it can maintain traces of student actions that can be used in improving problem-solving strategies, it can reify the process of thinking, not just the product, it can make the invisible visible, it can help create functional learning environments where the student can acquire goals and knowledge while pursuing goals that are meaningful to them, etc. Although his emphasis was on individual learning, many of his comments remain valid for our purposes. We see a number of these different uses of the computer discussed in this volume. At the most basic level, the computer can be used simply as a data-gathering tool that can support the investigation of collaborative learning processes between people, allowing for presentation of a task, and perhaps recording of responses, and later analysis of these responses. In this case, the computer makes the task of the researcher easier but does not really affect the collaborative learning process per se. The computer is playing a rather passive role that could be replaced, though more awkwardly, with other possible tools such as pencil and paper. To my mind, this does not really fall under the rubric of what CSCL implies. More innovative work provides a rich microworld on the computer which students can interact with individually and collaboratively (e.g. Behrend & Roschelle, this volume, O'Shea et al., this volume). Micro-worlds provide powerful settings for both individual and collaborative learning, especially allowing students to explore a world and develop their understanding, as witnessed in some of the work in this volume.
The computer can also be seen as a "tutor" with whom the pupil interacts, or even collaborates. One can discern two kinds of research activity in this direction. One, focusing on more traditional Computer-Aided Instruction (CAI) or Learning (CAL), in the form of "drill-and-practice" type exercises, is no longer regarded as having a major role to play in the classroom, and certainly not in the context of "collaborative learning". The other, Intelligent CAI, received a lot of attention in the previous decade, especially with the concept of the computer as tutor or coach that could help the student in understanding a problem domain by pointing out flaws in the students conceptual model evidenced by their responses to problems (e.g., Sleeman & Brown, 1982). While such a perspective does have more of a flavour of "collaboration" about it, between the student and the computer, the difficulties of building in an adequate set of strategies concerning how human tutors support students, and of understanding the students conceptual model of the domain have resulted in a major retrenchment of this work, and aspirations have become more mundane. The implicit assumption that the computer could in some sense replace the teacher is also now seen as flawed.
Yet a third viewpoint, and one I support, stresses the use of the computer as a medium or resource for collaborative learning (Crook, Newman, this volume). The computer can help students to communicate and collaborate on joint activities, providing assistance in the coordination process. This mediational role of the technology emphasizes the possibilities of using the computer not simply as an individual tool but as a medium through which individuals and groups can collaborate with others. In such studies the computer acts as a support and resource for the collaborating students. We will see some examples in the next section. These differing perspectives can affect what questions we ask in our empirical work, what experiments we conduct, and how we judge success. Again, the argument is not about which perspective is correct, but within which perspective is a particular research study conducted. Taking this into account can help in making a more coherent assessment of "success" or "failure" of the study.
2. CSCL Examples and Experiences
For several years now, I have been interested in how to enhance possibilities for people to communicate and cooperate with each other through use of computers. This includes the use of asynchronous computer-mediated communication applications such as electronic mail and computer conferencing, as well as more advanced facilities. In this Section I give some examples of systems and their uses in attempting to support collaborative learning, including some personal experiences.
2.1 Computer Communication Facilities
2.1.1 Uses of Electronic Mail
Basic electronic mail has been used for some time to provide a simple computer-mediated learning environment for students (e.g. Levin et al., 1984). Whether the computer plays any unique role, other than allowing for participants to communicate is questionable here, but the immediacy of the medium, and its ability to go beyond the classroom walls, even across countries, certainly does have an appeal. Also having a computerized trace of the activities engaged in by the students can be useful, as it allows teachers and students to study their exchanges and engage in new learning activities involving spelling, idea coherence, presentation, etc. It is important to note that the technology per se is usually not the crucial issue, rather the social practices surrounding its use. Simply providing a physical or electronic connection between people does not guarantee that any collaborative learning will take place. The important thing is to create a social activity through which learning can occur. Indeed, a nice example of how simple email exchanges can be upgraded to a more complex set of activities is given by Riel (1985) in the development of a Class Newspaper. This spawned a whole new set of activities such as editing, layout etc, which provided opportunities for the students to learn a variety of new skills.
While I have been involved in some successful experiments using email for collaborative learning, I have also had some failures. One concerned a hasty attempt to link some University students in Denmark and the USA, where the teachers in the 2 classes, myself and a colleague in California, were both lecturing on computer-mediated communication. In this case the initiative came from the teachers, and we had not really jointly constructed some useful context for learning to occur, i.e., no shared task that could be collaboratively accomplished by students in the 2 countries, other than a wish for the students to discuss this computer medium, while actually using it in this process. However, such a general and diffuse goal of exchanging messages about the medium was insufficient to maintain the dialogue between the groups. The lack of appreciation of what the communication was for, meant that after a short exchange, despite some efforts by the teachers to stimulate debate, the attempt at collaboration ended! There were a number of other mitigating circumstances in this case however. The backgrounds of the students were quite different. The students in Denmark were senior-level, and involved in Computer Science, the California students were sophomore, and in Communication Studies. The two groups of students did not have that much in common, as regards their computer experiences. Also while language was not a major barrier, fluency in the use of English varied among the Danish students, making some reluctant to use it, especially in a situation where the others involved were not known to them. There was also a more general cultural difference, in that people from this part of Denmark are known to be rather reserved in terms of social communication, while Californians have a somewhat different reputation! Despite such extenuating factors, I believe that these kinds of cultural barriers could have been overcome, and have been in other studies, when the need for collaboration and the nature of the joint activity is made clearly evident to the students. This requires that those involved in setting up the activity put in a lot of work to help structure it, especially in the early phases. Hopefully, once the project starts successfully, both students and teachers will "take-off" in a variety of directions, often developing quite new and unanticipated ideas and uses for the medium.
2.1.2 Uses of Computer Conferencing
With computer conferencing facilities, as distinct from email, some further support for collaborative learning is provided, as such facilities provide a transcript of the communication between students, allow for the spinoff of subgroups on various topics, and often provide shared access to material. This allows for some more coherence in keeping track of discussions and again has been shown useful to enable people to collaborate on accomplishing various tasks. Many studies of people using conferencing systems, such as EIES, COM, etc. have been done (see Kerr & Hiltz, 1982, for an early review of systems). There has been particular interest in the use of this medium as a vehicle for delivering distance education. Within some circles in the field of Collaborative Learning, this kind of use is regarded as being almost prototypical (Mason & Kaye, 1989). Kaye (this volume) discusses experiences in using computer-mediated communication on certain UK Open University courses, and compares different media in terms of user satisfaction, costs etc. While I believe that there is a use for the computer medium in this context, I am concerned about the way in which some studies make direct comparisons between educational courses, with the delivery vehicle or medium as the independent variable. A clearer example is Hiltz (1988), who summarizes a research project run in a quasi-experimental style comparing "virtual" classrooms where instruction and discussion all comes through the computer medium,with standard classroom instruction. To my mind, this kind of comparison is based on a flawed view of the role of technology in supporting collaborative learning. It tends to focus on the medium of instruction as if we can compare different media directly, i.e. without taking into account the nature of the instructor, the material, the students, etc. Even in studies where an attempt is made to control for some of these factors, trying to isolate the effects of any one technology in such a fashion reifies the technology at the expense of the social practices of the participants that make the system work. Arguments about substitutability of media based on this model are thus flawed, in my opinion. A more appropriate conceptual framework would seek to identify how certain computing environments might contribute to accomplishing activities, rather than viewing activities as being neatly decomposable into isolated tasks that can be allocated to students via different media.
Rather than viewing new media as replacements for others, we can construct contexts in which a variety of media can be used to support collaborative learning practices. I have used computer conferencing systems as an integral part of courses where traditional face-to-face interaction in the form of lectures and tutorial groups are also present. This allows for a richer set of activities within the group, new avenues which people can explore. Here, the focus is not one of substitution, but of augmentation of the methods and means through which learning can occur. One interesting finding was that the medium could provide opportunities for people perhaps inhibited in social situations to "speak out". One of the people who took little part in the face-to-face sessions became one of the most voluble on the computer medium. Over time, we were able to use this fact to re-integrate the person back into the face-to-face meetings as well, through reference to what had happened through the other medium. While initial activities on the medium were structured, it was quite gratifying to see how students appropriated the medium for their own purposes, setting up several conferences on their own, exploring the system, finding how to send anonymous messages etc. It is also interesting to note that these were the same students mentioned earlier who did not take to the idea of open-ended electronic mail exchanges with people in another country.
2.1.3 Open Bulletin Boards
An interesting example of the use of an electronic network for collaborative learning via information dissemination and general discussion is USENET, the worldwide Unix(TM) computer network. USENET supports mailing lists and bulletin board services concerning an amazingly large variety of work and social activities. It has some interesting properties and encompasses a very widely distributed user group. Of particular interest is the requests for information or help that can be sent out on the network, and the responses generated. Of course, some of the quick, and detailed responses that are characteristic of this medium can be explained by the novelty of the electronic medium, but this certainly will not suffice, as it has been in existence now for a number of years, and the phenomenon of quick and detailed responses to help requests is still apparent. Part of its appeal is that members believe themselves to belong to a community of fellow-travellers, -mainly software engineers and computer scientists, but basically open to all who seem willing to join in use of the medium, and respect some basic etiquette. An example of an extremely important, for some indispensable, means of support is evident in the unix-wizards group where system programmers ask for and often obtain extremely detailed help from others about the quirks of particular unix implementations. This service is indispensable, as often there is no one in the immediate vicinity of the person that could help with the problem. Being a member of this group gives one access to an incredible wealth of talent and expertise, at no cost either, other than accepting that at a later date, you should be willing to assist another in a similar fashion. Here is a case where a spirit of mutual support and cooperation has developed and persisted, even without any central organizational authority or large scale face-to-face meetings to encourage compliance with the "rules of the game" - although there are meetings of Unix user groups and they also serve important roles as information sharing venues and social support networks. While this may not immediately appear to be an example of CSCL, certainly people are learning, exchanging information, over time, through the computer system.
2.2 Local Area Networks (LANs)
Developments in technology, such as the increasing number of local area networks in many settings, including schools, provides new opportunities for CSCL. We start to see a number of interesting emergent phenomena that would be difficult to conceive of as developing without the computer medium. A nice example is that of Earth Lab (Newman & Goldman, 1987, Newman, 1988). Here students and teachers collaborate with other students and teachers in doing science, collecting observations, making experiments, and reporting their findings. The databases created provide new opportunities for sharing information and creating new activities and a variety of tasks that can be assigned, delegated, negotiated, and all mediated through the computer. The power and flexibility of the computer allows for the simultaneous coordination of multiple activity systems, for students and teachers performing a variety of interconnected and local tasks.
2.3. Hypermedia
The area of hypertext, multi-media, and hypermedia systems (Conklin, 1987) has generated a lot of interest in terms of its educational potential, and could be a potent tool in CSCL. Such applications can be simply defined as chunks of text (or other material - voice, graphics, pictures, etc) that are linked together in a network. This non-linear structure can be perused in a variety of ways on a computer system. Users could also add comments and notes to various items, which could be retrieved by others. The potential of hypertext-based collections of material could be a truly powerful resource for collaborative discovery and learning. I will briefly outline a small project in this area that generated some not unfamiliar problems when working with state-of-the-art systems.
We have been working with a powerful hypertext system - NoteCards (Halasz, Moran & Trigg, 1987). Our intention was to develop a pilot "community information base" for users of our computer workstations. One of our interests in developing this system was to develop what might be called the "community memory" of the environment - to encourage users to annotate the information within the database so as to make it more contextually based and meaningful for others in the work community. The possibility of critiquing the information, suggesting new links between existing features, examples of use of system features, guided tours (Trigg, 1988) of the database etc. were apparently easy to support technically in the Xerox NoteCards hypermedia system that we were using, but the organizational and social factors involved in promoting this activity were not, and were part of the reason for the investigation. We worked on ways to encourage users to contribute with information on experiences with the computer system through comments, criticisms, and general annotations on the documentation that had been provided. For this purpose we needed a medium such as hypermedia to support storage and retrieval of the loosely structured information in a shared database. The hope was to support the community in "growing" the database itself through the contributions of the various users, thus ensuring that the additional information provided was context-relevant, coming from the actual working practices of the user community, not from an external group.
In the course of the study, we found that much of our time was spent in overcoming a number of technical problems regarding poor response times, and interface issues, problems that we had not forseen would be so troublesome. While we eventually had a working prototype, the social-organizational issues involved in its use proved even more troublesome. We only had a few machines available to the students, as they were expensive workstations, and the course lasted a half-term, so it was insufficient time for the students to both get used to the system and "go critical" in terms of developing a sense of community among themselves, to really evolve the use of the system to the point which we were hoping for. We did learn from the experience, but our original goals were not met. A lesson that can be learned from this experience is that while powerful new applications may provide interesting new functionality, there is a still a big gap between performing studies based on a high performance development system and having a tool that is ready for use in an institutional setting, where cost, robustness, and reliability are crucial issues. The gap between what is possible in the lab and what can be expected to work in an ordinary educational setting is still large. However, we expect that in the next few years useful, robust hypermedia facilities will become more generally available and will offer new opportunities for CSCL in ordinary, everyday settings.
2.4 Further Developments
There are a large number of other technological developments, in areas such as - shared screen facilities, interactive video, groupware applications - that could be exploited in CSCL. Since a number of these developments are explicitly designed to support group or collaborative working through multi-user access, for example, it should not be difficult to develop interesting educational uses for such systems. Shared screen facilities, e.g. Timbuktu (TM Farallon) are available now for personal computers and could have immediate uses as collaborative instructional media. We have been experimenting with such applications to support group tutorials, where they have obvious benefits.
3. A Cautionary Note
While there are undoubtedly many and varied ways in which the computer can be used to enhance collaborative learning practices, and an ever-growing arsenal of computer tools that we can exploit in creative ways for CSCL, we should also be aware that this very plethora of tools can create problems. It is important to realise that our goal is to improve the conditions for learning through the use of technology. As noted already in the paper, there is a tendency to focus too much on features of the technology per se, and not on the learning activity. We must remember that social, organizational and political processes will always be involved in the adoption or rejection of particular technologies, both at the institutional level, and within the classroom - by teachers and students. As a recent review notes:"In the educational computing arena, the battle lines are clearly drawn between the revolutionary forces who would use the computer as a wedge to open up education to new ideas, new teaching techniques, and new goals for students, and those conservative forces who view the computer as a means of strengthening the existing curriculum rather than changing it. (Kurland & Kurland, 1987,p.326.) This underlines the need to be sceptical of claims about how computers are inherently "empowering".
Once a technology is in place, further problems may arise, as expressed in the mentality "If we've got, let's use it". This argument leads to ill-advised attempts to use computers to "support" learning activities, but it is not based on any clear idea of how joint activities will be supported through the technology. In these cases, it is quite possible that the introduction of the computer system may actually result in the disruption of an existing quite adequate collaborative process, rather than its support! This can be caused, for example, by moving a collaborative activity (that is quite adequately handled "offline") onto a computer system, resulting in teachers and students being lost in the details of trying to use a complex computer system - despite supposedly "user-friendly" interfaces - and unable to get to the stage of actually getting into, or through, the system to do the actual collaborative work! The computer can thus become an obstacle to be overcome, rather than a support for collaboration.
Conclusion
Let me summarize some of the key points raised here. A very basic issue for the field of CSCL is to come to terms with the multiplicity of contexts or environments in which people learn, and in which computers might play an important role. We have noted the wide variety of settings to which the term collaborative learning might be applied, from people engaged in a distance education course, to a pair of students collaboratively solving a problem that appears on a shared computer screen. We have also noted how the joint activity can vary from a few seconds to years, as when a group collaborate in performing experiments or in writing a large report. We have noted that learning is not confined to the classroom, indeed, we do not have to fully agree with the "de-schooling critics" such as Illich and Dennison to accept the fact that other settings for learning can be very effective. Nevertheless, given the current make-up of education in our society, an emphasis in the discussion of CSCL to-date on classroom activities is not surprising. This still allows for a wide variety of approaches however, in how the role of the computer in the classroom is envisaged. If we accept that the educational system cannot be changed overnight, then we need to understand more fully the current demands on teachers within the educational system in order to see how to best support them and their students.
This paper has argued that the best way to regard computers with respect to supporting collaborative learning is as an enabling medium through which the partners, whether teacher-pupil, pupil-pupil, teacher-teacher, or even larger scale group or organizational entities can organize and accomplish activities. Acknowledgement is given to the actual structure of classroom activities, and the computer tool is adjusted to try and fit in with on-going activities and teacher and student goals, rather than usurp them. The possibility of using the computer system to go outside the individual classroom activity - under development by a number of researchers such as Cole, Griffin and others - and provide access to other contexts of activity which can be mutually supportive provides an intriguing link between classroom activities and the activities of the outside world, breaking down some of the educational stereotypes that have held sway for too long.
In conclusion, let me re-iterate that the computer should be viewed as a support tool that can provide functional learning environments. When instruction is seen not simply as the imparting of information, but the creation of a space - Newman et al.'s "construction zone" - the power of the computer to allow both students, teachers and perhaps others outside this system, to reorganize their activities and develop in unexpected directions could be very significant. The computer would thus become an important adjunct to truly collaborative learning!
References
Cole, M. (1985). Personal communication.
Conklin, J. (1987) Hypertext: An introduction and survey. IEEE Computer, Sept 1987, 17-41.
Gagne, R.M. (1968) Learning hierarchies. Educational Psychologist, 6, 1-9.
Hiltz, S. (1984) Online Communities. Norwood, NJ: Ablex.
Oettinger, A. (1969) Run, Computer, Run - The Mythology of Educational Innovation.
Papert, S. (1980) Mindstorms. New York: Basic Books.
Sleeman, D. & Brown, J.S. (Eds.) (1982) Intelligent Tutoring Systems. New York: Academic Press.