Designing educational software as a re-usable resource

Journal of Computer Assisted Learning (1996) 12, 114-126

Designing educational software us a re-usable resource 1. Neilson and R. Thomas ICBL, Heriot Watt University and Department of Engineering, Cambridge University

Abstract The production of educational software is moving from a focus on courseware produced by the individual lecturer to the development of educational resources that can be shared and accessed from a wide variety of teaching sites. This change has a number of implications. Firstly, different resources must be capable of being seamlessly integrated into a specific teaching package. Secondly, collaboration between institutions will be increasingly important to ensure that resources so produced are sufficiently generic to be of interest to those outside the development team. Thirdly, management procedures to direct and control this collaborative effort are required. This paper details how the distributed hypermedia environment provided by the World Wide Web may be effectively exploited to realise these three goals.

h p d s : Hypermedia; Simulations; World Wide Web

Strategies for the production of educational software

The production of educational software is moving from a focus on courseware produced by the individual lecturer to the development of educational resources that can be shared and accessed from a wide variety of teaching sites. The model of the individual lecturer creating software is under attack for having resulted in the production of many items of courseware that are only used in the context in which they were developed and in the duplication of effort by individual lecturers in different institutions. If such duplication of effort is to be avoided it is realised that an "unprecedented degree of collaboration and sharing between different institutions" is required and that "new organisational structures will be needed in order to bring all this about" (MacFarlane, 1992). Avoidance of the inefficiencies of traditional courseware production is possible only if courseware production is viewed as the creation of reusable resources which address the core subject matter of a domain yet remain sufficiently flexible to permit tailoring to a particular lecturer's teaching style. "When technology is exploited, what

Accepted: 16 October 1995

Correspondence: I. Neilson, Department of Computer Sdence, Liverpool Universi Liverpool L69 3BX Email: ien%sc.liv.ac.uk 114

Designing educational software as a re-usable resource 115

should be created and used are resources which can be flexibly tailored to individual needs and used in a highly interactive way." (MacFarlane, 1992).' Similar points are made by researchers in other European countries (Collis & De Diana, 1994).

The consortia model of educational software development

One model which has been proposed for facilitating the collaboration required for a resource based approach to courseware production in higher education is that of development of educational software by consortia:

"Because of the importance of collaboration for greater efficiency of production of materials, and for the greater likelihood of widespread dissemination of them, any central funds made available for development of these materials should be given only to collaborative consortia." (Laurillard, 1993 p. 230)

The National Funding Council for Education in Britain has been urged to adopt such a model. The current (1993-1996) Teaching and Learning through Technology Programme (TLTP) required that all bids for courseware production involve at least three institutions as members of a consortium.

The consortia framework is premised on the belief that consultation and collaboration between academic experts across institutions will lead to the identification of those areas of commonalty in a domain which all teachers require while also identifying differences in both the contexts and the manner in which the core subject matter is used in different institutions. The requirements specification produced by this activity is argued to ensure that the software design process is centred on the production of a resource that is central to the teaching of a discipline yet which is sufficiently flexible to permit tailoring to particular teaching contexts. These requirements may be modified by further discussion during the design development process.

For the consortia model to work collaboration and consultation among a large variety of people is critical during the whole of the software development process. This poses a problem as there is no established methodology for cross-institutional, distant collaboration. Generally, it has not been subject to the discipline of software life cycle development methodologies or to the demands of team based collaborationt. There are no universally agreed mechanisms in place to help manage the process of collaboration.

* The MacFarlane report suggests the setting up of a Teaching and Learning Board which would create the necessary infrastructure from which to foster the widespread development of innovative teaching of all types, including the technology of education. It would produce a national strategic plan based on what the LJK higher education system should be like in 20 years time. The first five years of this plan is seen to involve the production of teaching materials by consortia (pp. 31-33). The exception to this rule is UK Open University ourses. There, the collective work of specialists gives the content of the teaching materials associated with these courses a special authority.

t

116 I. Neilson & R. Thomas

Two problems, one solution?

The problem of creating re-usable software and the problem of supporting collaboration across institutions during the design process are distinct. The former is concerned with the creation of an architecture for a software design that is sufficiently flexible to meet the requirements of lecturers wishing to tailor resources. The latter is concerned with supporting collaborative design at a distance through technology or indeed any other means.

This paper discusses how an architecture developed in the domain of engineering to solve the first problem, that of creating reusable software, may be used to facilitate the group decision making process that is at the heart of the consortia model for software production.

Background to the Interact Project*

The architecture we describe centres on the integration of the distributed hypermedia environment provided by the World Wide Web with simulation software. Both hypermedia and simulations are well recognised as effective learning tools. Hypermedia has been used to create effective, interactive exploratory environments (Nielson, 1990b). Both its advantages, the flexible structuring of course material that it permits, its facilitation of user defined browsing of educational material and its disadvantages, the inherent navigational problems that it can present to the user and the often fragmentary perception of the structure of a domain that a novice user can experience in using this material, are known (Nielson, 1990a; Perlman ef al., 1990). Similarly with simulations. The rich exploratory environment that a simulation can provide offers the student an experience of a domain that is more direct and therefore often more engaging that the 'second-order' description of that domain that characterises academic dialogue (Laurillard, 1993). However, without guidance, particularly in the early stages of exploring a domain the educational value of a simulation is reduced (Thomas & Neilson, in press).

The aim of Interact was to marry these two educational technologies and thus avoid the weaknesses of each medium in supporting the learning process when employed in isolationt. A resource based approach to education production cannot centre on the production of resources per se. Primarily, it must remain centred on the production of software environments that actively engage the learning process but which, in addition, are re-usable resources.

The TLTP INTERACT project is a consortium of three sites, Cambridge, Heriot-Watt and Strathclyde Universities, concerned with the development of interactive exploratory environments for the teaching of core domain knowledge within engineering. The aim of Interact was not, however, to create a total learning environment for students. Issues such as learner assessment fell outside the project's remit. Despite this, many of the features of the Interact Simulation Environment, such as the Snapshot and Replay facilities, could be readily adapted to an assessment rather than guidance function.

t


A resource based architecture for educational software in engineering

In engineering, core domain knowledge is often reflected in the form of mathematical expressions. Many apparently distinct engineering phenomena may be modelled by the one set of expressions'. For example, pressure transients occur in all fluid systems. The mathematical methods used to describe pressure surge, water-hammer or pressure transient propagation, are applicable to a wide variety of superficially dissimilar engineering contexts: off-shore oil platforms, hydroelectric schemes, water distribution networks, long-distance oil pipelines, aircraft fuelling systems, sewers and vent systems (Swaffield & Boldy, 1993).

If the mathematical expression, the model underlying a simulation, could be separated from the graphical expression of the model then it would be possible to use the one set of mathematical expressions to illustrate engineering phenomena such as the water hammer effect in a wide variety of contexts. If the simulation environment so developed could be linked flexibly to other media of expression such as text, graphics, video then the individual lecturer would be able to customise the use of the simulation according to the demands of a particular instructional context, be it lab, tutorial etc., the level and background knowledge of the students and in accordance with a preferred teaching style. If the hypermedia environment to which the simulation was linked was distributed, then the resources available for instruction and learning would be increased enormously. Re- use and augmentation of any created resources would be facilitated.

Such an environment has been created within the TLTP project, Interact. The basic architecture of this environment, the Interact Simulation Environment (ISE), is shown in Fig. 1. Realisation of such an environment has been made possible by the development of the World Wide Web and by the creation of two tools, the Interact Communication Facility (ICF) (Slater, 1993; Slater, 1994a; Slater, 1994b) and the Presentation Manager (PM) (Smeaton & Slater, 1994) that exploit the functionality offered by the Web.

In the Interact context 'core commonality in teaching content' is reflected in the design and choice of the generic model, the set of mathematical expressions, that is at the core of the simulation environment. The model is expressed through an Object Oriented Graphical Interface (The Presentation Manager), the presentational context of which may be altered by an individual lecturer or user. This allows the lecturer to illustrate the operation of the model in a variety of contexts. It also allows the individual lecturer control over how a given engineering context is displayed to the student. The PM also allows the state of a simulation at a given moment in time to be captured, saved to file and reloaded at later date (Snapshot facility). The file may be reloaded locally or accessed over the Web by a remote user thus permitting simulation states to be shared and communicated about at a distance. A sequence of interactions with the

* A different and more complex example is in the use of Laplace's equation in engineering where it can be used to model ground water flow, electrical fields or heat. In this case, the equations model physically distinct effects rather than the same effect in different contexts.


simulation may similarly be recorded and saved to file (Replay facility). The latter facility allows the student to reflect upon past actions.

_ _ - _ - - - - - - Text images,

Mosaic and the World Wide Web

/ 0

0 /

0 0

I-

- -

\ \

_ _ _ _ _ _ _ _ - - Messqe + passing

Interact Communication Facility - _ _ _ _ _ - - - -

Simulal

Model (Computational

Resentation hlanager (Students see this as irnphiul user interface) . - - - - - - - - - ion

Fig. 1. The structure of the Interact Simulation Environment

The simulation is linked to the surrounding hypermedia environment provided by a standard Web browser, such as Mosaic or Netscape, by the Interact Communication Facility (ICF). The ICF permits simulations to be controlled from scripts embedded as links within HTML documents and, reciprocally, HTML documents to be controlled from within the simulation itself'. Help documentation, written in HTML, can therefore be accessed from within a simulation.

Use of the simulation in different educational contexts, laboratory, demonstration, lecture, self-study etc. is accommodated as the form of any accompanying instructional material is determined by the individual lecturer authoring in HTML. Different levels of guidance in use of the simulation for different levels of students can be provided. Flexibility in the purpose to which the simulation is put and individuality in teaching style is thus achieved.

The basic power of the ISE environment comes from the treatment of simulations as but a mode of expression that should be able to combine freely with other, more familiar, modes of expression, such as text, video, sound etc. to yield a particular items of courseware.

The generative nature of the ISE is illustrated in Fig. 2 below. Four different ways in which the one simulation might be used are illustrated: the simulation is used by lecturer A as a stand-alone demonstration in a lecture with an accompanying video (use 1). The same simulation is then integrated with appropriate textual material for use in self study and revision. Pre- recorded simulation states are called from the textual material using the Snapshot facility provided by the Presentation Manager (use 2). Lecturer B embeds the simulation within her lecture material and uses a different video.

The ICF, as currently implemented, involves no modification of these standard Web browsers. Full technical details concerning the ICF may be found in Slater, (1994) and also accessed through the World Wide Web (see end reference section).

Designing educational software as a reusable resource 119

She also modifies the interface to the simulation to remove some of the more complicated options as her students are less advanced than A's (use31 ; Lecturer B produces a complete course for more advanced students and tailors the interface to the simulation differently for different parts of the course to focus the students' attention on the salient issues (use 4).

Authoring

Different items of courseware

Fig. 2. The generative nature of the Interact Simulation Environment

Benefits of a resource based design

The flexibility afforded by the ISE is extremely powerful. The open nature of the system design permits:

The environment to be readily linked to any other educational resources considered important in the total context in which the learning is taking place e.g. sources dealing with the necessary background knowledge for understanding the model on which the simulation is based, references to additional reading, information about other knowledge required to use the ISE effectively e.g. knowledge about windowing environments etc. Greater control by the lecturer over which aspects of the core domain knowledge embedded within a simulation are expressed within a given educational context. The interface to a model may be edited to exclude more complex options thereby simplifymg the learning task facing introductory level students. Even the manner in which a particular


object, for example an oscilloscope, is graphically portrayed may be controlled by the individual lecturer. Indeed, if desired, the simulation may be controlled completely from within a hypertext tutorial document written by the lecturer. This allows the lecturer to exert greater control over how the simulation is used. It allows the student's attention to be centred solely on solving the domain problem. Such a facility might prove particularly useful with computer naive students or when a tutor only wants to briefly reference a simulation during the delivery of a course. Obviously, student control of the simulation would remain important in other contexts such as exploratory learning. The seamless integration of email facilities within the environment. Not only can Mosaic's fill-out form facility be used to enter data intended for applications within HTML documents, as in the example of calling simulations from within tutorial documents, but it may also be used to enable students to send queries about a course to a lecturer. When running a simulation the user is also provided with a snapshot facility which can save the current state of the simulation to a file at any time. The data is automatically saved in the Interact Communication Language (ICL) so that it can be controlled over the Web. This is an important facility as it greatly extends the nature of the communication that can take place between lecturer and student, regardless of distance. Not only can the student readily email the lecturer about a problem from within the ISE environment but s/he can include the http address of the HTML document that references the file which encodes the values of the simulation variables that describe the problem state. By loading this file over the Web into a locally running version of the simulation, the lecturer can see the student's problem. Reciprocally, the lecturer can include links to simulation states in hidher response to the student. Many of the problems associated with text only discussions of student' difficulties are thus avoided. A mail routing mechanism may be used to direct students' queries to appropriate domain experts when more than one site or course is involved in a teaching exercise. The environment thus begins to provide the student with discussion level tools (Cumming 1993; Cumming & Self, 1990). These exchanges between student and lecturer can be kept and used to construct an ejectronic database of common questions and answers, an Answer Web (Smeaton & Neilson, 1995). This database can be structured in various ways, including by reference to the simulation states that gave rise to the questions. This has potential value as an educational resource for use by subsequent years of students. An electronic record of solutions to common problems is likely to prove important as the numbers of students engaging in study without ready access to a human tutor increases (Mayes & Neilson, 1995). A record of the type of questions students raise when engaged in using the software is also important for evaluation purposes. Such questions provide useful

The Answer Web idea originated from the work on the Answer Garden by Ackerman and Malone (Ackerman & Malone, 1990) on organisational memory.


feedback on the current educational effectiveness and usability of the courseware which can be used to further improve the design of the courseware (Neilson et al., 1993). Finally flexibility in structuring is not only an important facility for the individual lecturer it also is a facility that may be exploited by a particular student. If, as is frequently claimed in the literature, teaching a subject is one of the best ways of learning it then the provision of flexible architectures of the type described above is not only important to lecturers but also to the student as s/he may equally be seen in the authoring role.

Facilitating the design development process through the Web

We have thus seen how the integration of simulations into a the distributed hypermedia environment provided by the World Wide Web through use of the Presentation Manager and the Interact Communications facility can be used to create reusable software for engineering education. Can these facilities be used to facilitate the group decision making process that is at the heart of the consortia model for software production? While the production of tools for the latter purpose was never an expressed goal of the Interact project, the PM and ICF tools have been found to provide useful facilities for improving communication at a distance about design ideas during the software develoument life cycle.

In the same' way that &dents are able to incorporate snapshots of the actual problems they encountered in running a simulation in a communication to a lecturer, so too, members of the software production team can incorporate snapshots of simulation states into on-line design evaluation reports accessible from any site through the Web thus facilitating communication about problematic aspects of the design of prototypes at times which suit the individuals involved. The problems in scheduling design meetings with academic staff are bypassed. The Presentation Manager permits suggestions for new screen layouts generated during the design development process to be saved to file', accessed over the Web and reloaded into the simulation environment and tested at another site. Communication and the discussion of ideas between programmers and domain experts alike is thus facilitated. The rapid prototyping and sharing of ideas, so important to successful design work, is maintained despite the distributed nature of the software team. Further, the ease with which a proposed layout for a screen design may be changed reduces the expert knowledge required to communicate design ideas and thus facilitates the adoption of a participative design methodology. Rather than having to rely on verbal expressions for communicating their wishes, domain experts and prospective student users alike can readily alter a proposed interface object themselves, save it to a file which is referenced in an associated, automatically generated

* Interaction objects, designed by the simulation toolkit, are created with an automatic ability to save themselves in ICL format.


hypertext document. The URL' of the latter may then be emailed to the programmer. By simply calling the URL over the Web and clicking in the document hotspot, the programmer causes the data referenced by the document to be automatically fetched from the remote site and loaded into the simulation that he is running locally. The alternative design idea is thus directly viewed. The whole communication process takes only a few minutes. The form fill out facility provided by Web browsers such as Mosaic or Netscape may be exploited in the service of design as well. Whenever a new graphical object is created, its existence may be notified to all members of the design team by the automatic generation of a form which contains the details of the object's specifications. A database of such forms may be created and used to facilitate both bug control and the reusability of software. As Nielson (Nielson, 1989; 1990b) has pointed out, hypertext, (and thus hypermedia), can be used for storyboarding an initial prototype design. Within Interact a tool developed to assist use navigation through Web documentation, Footsteps (Nicol et aI., 19951, has been used to facilitate the construction of such on-line storyboards. Such storyboards have the advantage of being accessible to academic experts at physically remote sites for cross-expert evaluation.

Managing the design process in a distributed hypermedia environment

For the power of Presentation Manager and the Interact Communication Facility to be harnessed in the service of good design, the increased flexibility in communication that they offer has to be matched by a management structure which keeps track of the disparate, interweaving communications among members of the group. This structure has to ensure that all members of the group are kept informed about discussions, both on-line and off line, that have implications for design decisions. It has to facilitate the extraction of common problems from design discussions occumng in different contexts and permit the generalisation of solutions generated in one context to that of another where appropriate. Most critically it has to foster collective responsibility for the design process as a whole despite the fact that members of the design teams are situated on different sites with no history of co- operation and no institutional structure for support of the collaborative group decision making process which is at the heart of the design by consortia model.

Collective responsibility is not something that an electronic form of communication by itself will facilitatet. However the Web does provide an extremely flexible as well as readily accessible facility for group

*

t

URL stands for Universal Resource Locator. It is the means by which documents placed on the Web may be found. In Interact we continue to use face to face meetings in the form of workshops and video conferences to engender a group spirit and provide a platform for the sharing of expertise.


communication*. It is our argument that the exploitation of this in the service of all the stages of the traditional software development life cycle will provide the necessary infrastructure for group decision making and thus promote feelings of collective responsibility for a process of design that is not rooted in any given physical location. By requiring that all documentation for each of the stages in software development be written or translated into HTML, placed on the Web and appropriately linked, a structure is created that allows all members of the development team to have access to the basis of any design rational. Proposed solutions to design problems can be checked against initial requirements specification documentation. Requests for action as a consequence of design decisions can be posted and their status readily reviewed at appropriate intervals by members of the design team and the project manager. This can help ensure a momentum to the design processt. Transcripts of design meetings* may be placed on-line and linked to appropriate sections of the software development documentation thus permitting those not physically present at the meeting to have readily review the issues discussed. Email discussions of design issues have likewise been converted to HTML and linked to appropriate other documentation.

Issues that require to be addressed

We are aware that this attempt to use the Web as a tool for the management of collaborative design is raising many issues that have been raised in discussions of CSCW work, (notably that of Conklin and Begeman (1988) and in the general literature on hypermedia support for software development (Roth et al., 1994). Many commercially available hypertext systems for the management of collaborative design are considerably more sophisticated by way of facilities for version control of reports, documentation (Delisle & Schwartz, 1987), for differentiation of the structure of the hypermedia space by node type (Conklin & Begeman, 1988) and provide considerably more aids for authoring than that provided by public domain browsers such as Mosaic (Shneiderman, 1989). However this situation is changing rapidly§ and even if it were not, the distributed nature

Through the Web itself we have become aware of one other project that is using it in this fashion, that of the MADEFAST project (URL: http: //cdr.stanford.edu/html/ MADEFAST/home.html). MADEFAST Is an experiment in collaborative engineering, conducted over the Web. However, a cautionary note. The work involved in the maintenance of the hypertext record of the design development process can be considerable. Programmers are often reluctant to take responsibility for the maintenance of their particular part of project documentation. This type of record keeping often seems to be regarded as a tool of management rather than an integral aspect of design activity (Grudin, 1991). Only where such record keeping clearly served the programmers own interest, as in the case the publidsing and recording of new interface objects, was a willingness to engage in this type of record keeping noted, Within OUT own project audio recordings are made of any informal or formal design meeting which occur. Critical design issues are extracted from the verbal protocol and placed online for all members of the development team to consider. For a review of recent Web based work on supporting collaboration, see http://www.w3.org under collaboration.


of the Web and the integrated set of tools that are available within the one environment far outweigh the limitations of the browsers per se'.

Indeed, there has been exponential growth of software for exploiting the facilities offered by the Web. One of the new software techniques, that of mobile codet, offers a potentially more sophisticated solution than that provided by the ISE to the problem of how to integrate applications such as simulations into a hypermedia environment.

Mobile code is code that can be transmitted across the network and executed at the other end. An example of such code is Sun's Java language. Simulations written in Java could be transmitted across the network and downloaded as required by Sun's HotJava browser. Use of mobile code would avoid the necessity of the user having to store a copy of the binary of the simulation locally, as is required by the ISE. Rather the simulation could be downloaded dynamically as required through the browser. This has the additional advantage of ensuring that user has the most recent version of the software. More critically, however, Java as a language is designed to be platform independent. Unlike the simulations produced using the PM which are Unix dependent, simulations written in Java would be capable of being run on a wide variety of different machines without further modification.

Interact has already developed a Fast Fourier Transformation simulation in Java and is currently exploring the extensions to the ISE that would be made possible through re-implementation in Java. However, few applications are as yet Java based and while that remains the case the functionality provided by the ICF and ICF type toolst will remain useful.

Summary

The distributed hypermedia environment provided by the World Wide Web is shown to provide a useful framework for the adoption of a resource based approach to educational software development. Two tools, the Presentation Manager and the Interact Communication Facility, developed by the Interact Project, extend the functionality of this environment to include simulations and in addition offer a language facility that permits the sending of any object written in the ICL format over the Web. This facility both extends communication at a distance between student and lecturer and between

The lack of integration between stand-alone hypertext programs and the rest of the user's computational environment has been a major source of concern to users in the past (Leggett, et al., 1990). Further, as was illustrated by the ISE, many of the basic features provided by Mosaic such as the form fill in facility can be adapted to other uses, for example, through use of the software tools provided by the Presentation Manager and the communication faduty provided by the ICF. For a full technical discussion of the various mobile code systems that are available currently, see http://www.w3.org under Mobile Code. Such as the MShell utility developed by Chand and Thomas (pers cum) which mimics some of the features of the ISE that is provided by the ICF. MShell permits Matlab to be started and stopped and to have commands sent to it from a Web browser such as Mosaic or Netscape. The state of a model may be saved and set and information about error states or other predefined conditions within a Matlab program can be displayed in a Web page.


members of the software design team. Both simulation states and interface designs may be accessed over the Web. Finally the hypermedia environment provided by the Web can be used to facilitate the management of the collaborative design effort. All the specification and implementation documentation that is produced during the software development life cycle can be placed on the Web, interlinked and accessed from all the sites participating in a consortium, regardless of physical location.

Acknowledgements

We gratefully acknowledge the contribution of co-researchers in the Interact project to the development of the ideas presented in this paper. In particular, we acknowledge the work of Alan Slater on the Interact Communication Facility and Glum Smeaton on the Presentation Manager.

Information about this paper available on Web servers:

ICBL: http:/ /www.icbl.hw.ac.uk

The ICF facility: http://viper.cee.hw.ac.uk/-afs/www94/www94.html

The Interact project: http:/ /medusa.eng.cam.ac.uk/-interact/ for a guided tour

The Simulation toolkit: http://viper.cee.hw.ac.uk/-calm/ or

http:/ /gorgon.cam.ac.uk/-CAL'95/

http:/ / www.ncsa.uiuc.edu/SDG/IT94/Proceedings/ Educ/smeaton/ise-www/ise-www94.html

The Answer Web: http: / / www.cee.hw.ac.uk/ -calum/AnswerWeb/

References

Ackerman, M.S. & Malone, T.W. (1990) A tool for growing organisational memory. ACM Conference on Office Information Systems, pp 31-39.

Collis, B.A. & De Diana, I.P.F. (1994) Portability and networked learning environments. Journal of Computer Assisted Learning, 10,2,125-136.

Conklin, J. & Begeman, M.L. (1988) gIBIS: A hypertext tool for exploratory policy discussion. ACM Transactions Office Information Systems, 6,4,303-331.

Cumming, G. (1993) A perspective on learning for Intelligent Educational Systems. lournal of Computer Assisted Learning, 9,4,229-238.

Cumming, G. & Self, J. (1990) Intelligent Educational Systems: Identifying and decoupling the conversation levels. Instructional Science, 19,1,1-17.

Delisle, N. & Schwartz, M. (1987) Contexts: A partitioning concept for hypertext. ACM Transactions Office Information Systems, pp 168-186.

Grudin, J. (1991) Obstacles to user involvement in software development, with implications for CSCW. International Journal of Man- Machine Studies, 34, 3, 435- 452.

Laurillard, D. (1993) Rethinking University Teaching, Routledge, London. Leggett, J., Schnase, J.L. & Kacmar, C.J. (1990) Hypertext and Learning. Designing

Hypermedia for Learning (eds. D.H. Jonassen & H. Mandl) Ch. 2. Hypertext Springer-Verlag, Heidelberg, West Germany.


MacFarlane, A.G.J. (1992) Teaching and Learning in an expanding higher education system. ISBN 0 9519377 1 5, Available from SCFC, POBox 142, Holyrood Road, Edinburgh EH8 8AH.

MAGPIE (1994) The Interact Educational Resources Guide. University of Wales, Aberystwyth. http://www.dcs.aber.ac.uk/-jjwO/index-ht.htm1,

Mayes, J.T. & Neilson, I.E. (1995) Learning from other people's dialogues: questions about computer based answers. In Innovative adult learning with innovative technologies (eds. B. Collis & G. Davies). Elsevier/North Holland, Amsterdam.

Neilson, I., Mayes, T. & King, I. (1994) Evaluating engineering courseware: an example of user centred design. IEE Colloquium on Computer Based Learning in Engineering, Institute of Electrical Engineers, Savoy Place, London.

Nicol, D., Smeaton, C. & Slater, A. (1995) Footsteps: Trail-blazing the Web. Third International World Wide Web Conference, Darmstadt, Germany. http:// ww.eee.strath.ac.uk/-nicol/footsteps/footsteps.html.

Nielson, J. (1989) Prototyping user interfaces using an objectaiented hypertext programming system. NordDATA'89 Joint Scandinavian Computer Conference, Copenhagen, pp. 485-490.

Nielson, J. (1990a) The art of navigating through hypertext. Communications of the

Nielson, J. (1990b) Hypertext and Hypermedia. Academic Press, London. Perlman, G., Egan, D., Ehrlich, S., Marchionini, G., Nielson, J. & Shneiderman, B.

(1990) Evaluating hypermedia systems. Proceedings ACM CHI'90 Conference Human Factors in Computing Systems. Seattle, WA.

Roth, T., Aiken, P. & Hobbs, S. (1994) Hypermedia Support for software development: a retrospective analysis. Hypermedia, 6,3,149-173.

Shneiderman, B. (1989) Reflections on authoring, editing and managing hypertext. The Society of Text (ed. E. Barrett) pp. 115-131. MlT Press, Cambridge, MA.

Slater, A.F. (1993) An Overview of Communication within the Interact system. Interact project, Heriot Watt University, Edinburgh.

Slater, A.F. (1994) Controlled by the Web. First International Conference on the World Wide Web, Geneva. http:/ /www.cee.hw.ac.uk/-afs/www94/www94.html.

Slater, A.F. (1995) Extending W3 Clients. Second International World Wide Web Conference: Mosaic and the Web, Chicago. http: / / www.ncsa.uiuc.edu/SDG/IT94/Proceedings/ DDay/slater/ew3c.html

Smeaton, C. & Neilson, I. (2995) The Answer Web. Refereed poster presentation at the Fourth International World Wide Web Conference, Boston. http: / / www.cee.hw.ac.uk/ -calm/ Answerweb/ paper.htm1

Smeaton, C. & Slater, A. (1995) Integrating simulations with W3 wursewure. Second International World Wide Web conference: Mosaic and the Web., Chicago, http:/ /ww.ncsa.uiuc.edu/SDG/IT94/Proceedings/Educ/smeaton/ise~www/ ise-www94.html

Swaffield, J.A. & Boldy, A.P. (1993) Pressure Surge in Pipe and Duct Systems. Avebury Technical, Ashgate Publishing Limited, Aldershot, Hampshire, England.

Thomas, R. & Neilson, I. (in press) Harnessing simulations in the service of education: The Interact Simulation Environment. Computers and Education.

ACM, 33,3,296-310.

Documents

Designing educational software as a re-usable resource