ACE publications : Published 1994 - Copyright the authors. All rights reserved.

Towards a Framework for Open Courseware

Phil SiviterDouglas Siviter
Department of ComputingSchool of Computing, Information Systems and Mathematics
University of BrightonSouth Bank University
Lewes RoadBorough Road
Brighton BN2 4GJLondon SE1 0AA
EMail : Phil.Siviter@brighton.ac.uk                EMail : D.Siviter@sbu.ac.uk

Abstract

This paper describes work in progress in one aspect of a programme of research into the design, implementation and management of computer based courseware. We place this paper in the context of this research programme, and of courseware development efforts within the UK and abroad. It is our belief that the current generation of courseware may not be exploited as widely as is hoped, due to inherent limitations in its design and implementation. We examine these limitations and their causes, and claim that the courseware development community desperately needs to adopt an open courseware framework for future developments. We use a scenario of a lecturer creating a tutorial from disparate sources of courseware to illustrate the benefits which open courseware would bring, and to derive some requirements for an open courseware framework. A foundation of such a framework is outlined at an architectural level, and some of the high level protocols required by the framework are presented. We also present a vocabulary and a logical structure, derived from the HyperCourseware reference model, which courseware objects can map their internal structures to, to become framework compliant. Finally we briefly describe some tools which facilitate creation, maintenance and management of open courseware objects, and outline areas for future research.

Keywords

HyperCourseware, Courseware Management Systems, Course Browser, Computer Based Education, Computer Assisted Learning, Computer Supported Collaborative Learning, CSCL, Courseware Development, Courseware Specification, Courseware Objects, EPOC, Standards for Open Courseware, Framework for Open Courseware, Courseware Protocols.

Contents


1 Introduction

This paper is part of a programme of research into a range of issues concerning computer based courseware. These issues include:

These are all issues which tend to influence discussions at all stages of courseware development. For example, discussions about template design will require consideration of all the above issues, as will discussions about courseware development tools and techniques.

Discussions of some of the issues of structure, views, navigation and orientation can be found in [LINE94, SIVI93, SIVI94]. Discussions of instructional strategies can be found in [BUND94]. This paper is specifically concerned with the issues of structural interfaces, resource interoperability and integration.


2 Context

The last few years have seen a significant increase in activity in the development of computer based courseware (henceforth referred to simply as courseware). In the UK, there are currently substantial investments being made in courseware development, in particular the Teaching and Learning Technology Programme (TLTP) funded by the Higher Education Funding Councils (HEFCs), under which projects are funded to develop ways of using technology to enhance learning in the UK Higher Education sector [CTIS93a]. This programme is a response to the growing recognition of the need to satisfy an increase in student numbers of 50% by the year 2000, without a commensurate increase in resources [TRAI92]. Phase one of this programme began in August 1992, in which 43 projects were funded across the UK. Partly in response to the success of phase one, the HEFCs have funded a second "batch" of 33 projects under TLTP phase two [CTIS93b].

This rise in interest is not confined to the UK. In other countries there exist thriving and enthusiastic communities of educators and researchers involved in the production of courseware resources of various kinds in a diverse range of subjects [HYPE94]. The international nature of these developments is witnessed by the growing number of international conferences on the subject.

The motives behind these developments are varied. Some of the stated aims include:

Whatever the hopes of the CAL community, the ultimate success of these ventures relies upon the resulting courseware being eagerly adopted and exploited by colleagues in the courseware authors' institutions and elsewhere. Unfortunately, there are many reasons why this may not happen [LAUR93]. The bottlenecks which impair the successful exploitation of courseware range from cultural to technical and are quite diverse. Within this paper we restrict our discussion to bottlenecks of a technical or design nature.


3 Limitations of the current generation of courseware

The courseware being developed today is for the most part strictly stand alone. It usually runs on one platform only, rarely imports or exports data, and is generally monolithic in its implementation, by which we mean that it is impossible to use a part of the courseware package independently of the package as a whole. However, courseware is an educational resource, and such resources share the following features (amongst others):-

As a result, a courseware package which may only be used in its entirety will be used by only a few lecturers (and only to the extent that they approve of its instructional approach). Moreover, even those lecturers who use it when it is first available will stop using it as soon as it goes out of date. We cannot prevent courseware from going out of date - subjects change. Courseware development projects should therefore address seriously the task of maintenance. Unfortunately, few projects do so. Updating even a small part of today's courseware packages can be inordinately expensive, due in large part to the monolithic nature of their design and implementation.

At this point it is useful to note the way in which lecturers construct courses using traditional materials such as books and journal articles. It is a rare delight indeed to find a text book upon which one can base an entire course. It is more likely that we will tell our students to "read chapter A of book X, chapter B of book Y, paper P (pages 10 - 15) from Journal J" and so on. We pick and choose, mix and match, from the available resources. This approach deals with the features of educational resources stated above. Textual materials can be accessed at many levels of granularity with equal (and minimal) effort, so it does not matter that any given book is not exactly what the lecturer requires. Moreover, evolving subjects generate "small scale" texts in the form of journal papers etc. which lecturers incorporate into their courses.

We need to be able to adopt a similar approach to the use of courseware. i.e., we need to be able to easily put together a package of courseware which enables the learners to (for example) work through simulation S from Smith's "Fluid Dynamics 2" package, then do self assessment exercises 4 and 5 on screen 3 of the package "Fluid Dynamics Made Easy" by E.P.Jones, and so on. Unfortunately, the current state of courseware technology is such that it is simply not a practical proposition for most lecturers to take courseware packages apart and reassemble them in this way. The next section explores some of the reasons for these limitations.


4 Courseware development - lessons to be learned from software engineering

4.1 Recognising the problem

The courseware development efforts of today can be compared with the commercial software industry of a few years ago. Until recently, software applications lived a life of isolation - word processors could not share files with other word processors, much less with a spreadsheet or a database, even if the applications were all written to run on the same operating platform. Today the inexorable move towards open systems has led to a state of affairs where most word processors can read (and often write) files in most other word processor formats. Spreadsheets and databases can import and export their data in a variety of formats so that data sharing between applications is almost taken for granted by users. More importantly, technologies such as Microsoft's Object Linking and Embedding (OLE), IBM/Apple's OpenDoc, and their supporting object models (Common Object Model (COM) and Distributed System Object Model (DSOM)) are enabling more dynamic interoperability of disparate applications by providing protocols for the import/export of data and functionality, not only between applications but also across operating platforms. This means that (for example) an OLE client application can embed a document exported by an OLE server application (even if the client was written before the server was even conceived of) and the user can invoke the functionality of the server application from within the client. The arrival (and enthusiastic adoption) of Visual Basic custom controls (VBX) and the pending arrival of Microsoft's OLE custom controls (OCX) is enabling the construction of applications by the straight forward assembly of powerful, reusable components.

In contrast, this open systems mentality is conspicuous within the courseware development community only by its absence.

The lack of interoperability may well have been an annoyance and frustration for users of standard software applications, but a similar isolation within the world of courseware has far more disastrous consequences. Cynics of computer based courseware enjoy pointing out how much money has been spent in past initiatives on the development of courseware which has subsequently been used either sparingly or not at all. The reason why "closed" software is more debilitating in courseware than in the more standard types of software application is that the two types of software are very different in nature. Word processors, for example, were very useful even before they could communicate with other software. Once you had one you could use it for years and still extract useful functionality from it. On the other hand, as we saw in section three, lecturers must be able to integrate small units of courseware from different sources or they are unlikely to use them at all.

4.2 Recognising the solution

It is our belief that the answer to some of the problems of maintenance and flexibility of use of courseware lies in the development of a framework for "Open Courseware Systems". Courseware conforming to such a framework would be capable of exporting its educational material, at varying degrees of granularity, to other software entities in its environment, provided that they too conformed to the framework. The following are some of the benefits which we believe would follow from the adoption of such a framework.

The following section briefly describes a scenario which illustrates the first of these points. It is the kind of scenario which we believe will become commonplace once an open courseware framework has become accepted. We also believe that this is not only a desirable state of affairs to aim for, but that it is actually a necessary pre-requisite for the large scale exploitation of computer based learning software.


5 Working with an open courseware environment - a typical scenario

5.1 Background

This is a scenario which may well be carried out on a "day before the tutorial" basis. A lecturer is teaching a course on Object Oriented Design, and is putting together a computer based tutorial for one of the topics in that course (e.g., the Dynamic Model from the Object Modelling Technique (OMT) [RUMB91]). Her department is well equipped with large amounts of computer based courseware stored on a central fileserver and accessible via a network at any of the student workstations on the university campus.

Among the available on-line courseware resources are Computer Aided Learning (CAL) modules covering a huge range of topics, including some specifically written for teaching aspects of:-

There are also many modules available for subjects as diverse as Fluid dynamics, Cosmology, Food Hygiene, Business French, and so on.

The lecturer has seen some of this courseware (but not all) so she has some idea of the kind of material which is available. Her intention is to browse through the on-line courseware and select a few items which she thinks will be suitable for her particular group of students, studying this particular topic (the Dynamic model from OMT) at the particular level of this unit (e.g., introductory). Of course, it would be very convenient if there was a single CAL module which met her requirements, but, just as it is rare to find a book which precisely meets a lecturer's requirements, so it is rare to find a CAL module which is perfectly matched to the objectives of a particular tutorial. Instead the lecturer will most likely have to use pieces of courseware taken from a number of different CAL modules.

The lecturer pieces together her tutorial using a software tool which we will call a "browse and build" tool. This tool interacts transparently with a number of other software entities to search for CAL modules, present lists of them to the lecturer, execute them on demand, and create a tutorial consisting of selected modules, or parts of modules. The resulting tutorial takes the form of a file to which the lecturer can direct her students, and which will result in the selected CAL material being presented to the students.

5.2 The lecturer's view

The lecturer logs in at her workstation, opens a window labelled "Courseware" and double clicks on the "browse and build" icon to launch the browse and build tool. This tool offers a number of avenues for browsing the existing collection of on-line courseware resources; e.g., it offers:-

Using the various search and import facilities of the browse and build tool, the lecturer finds a number of modules relevant to her tutorial. (We use the term "modules" here to refer to pieces of courseware of any size, from a single activity covering a single concept and lasting just a few minutes, to an entire course intended to be covered in the duration of a semester.) She reviews several Topics (i.e., she works through the courseware) in some cases deciding to use an entire Topic as it stands, and in other cases deciding to use some of the Activities from a Topic but ignoring others. After about an hour she has collected the following set of courseware resources.

From the Course "Object Oriented Design" she has an entire Topic "Dynamic behaviour" consisting of several Activities. From the same Course, she also has two Activities ("Object diagrams" and "Interaction diagrams") from the Topic "Notations for dynamic behaviour". Finally from the Course "Real Time Systems" she has the Topic "State transition diagrams", again consisting of several Activities.

To complete her tutorial she now needs to provide a convenient way for her students to access these modules. The browse and build tool has facilities for building simple maps, so she creates a blank map and adds four buttons, which she labels "Dynamic behaviour", "Object diagrams", "Interaction diagrams" and "State transition diagrams". She then selects the "Dynamic behaviour" button and chooses "Link to..." from the menu, which brings up a floating palette with buttons "Link" and "Cancel". She then launches the lessons map for the Topic "Dynamic behaviour" and presses the Link button on the palette. She is returned to the browse and build tool. Pressing the "Dynamic behaviour" button will now launch the lessons map for the Topic "Dynamic behaviour". She links the other buttons to their respective Topics and Activities in the same way, and then adds a graphics background to her map to make it more attractive. It is the background she always uses, and which her students now recognise as hers. Finally she chooses "Make tutorial file" from the menu of the browse and build tool. This creates an executable file which students can run, and which will present her new map.

The whole process has taken approximately one hour, most of which was spent reviewing the CAL material itself before deciding whether or not to include it in the tutorial.


6 A framework for open courseware

The previous section outlined a scenario of the kind which we believe will be commonplace within the next few years. It is our belief that in order to enable this kind of scenario we require an open courseware framework - an architecture and set of protocols which enable conformant courseware to exchange resources and functionality. Such a framework is currently under development by the authors. This section gives an overview of the architecture of this framework, while the following section outlines a subset of the protocols which have been identified.

Our open courseware framework is based on communicating/co-operating courseware objects. We distinguish between two views of the framework. The first view, which we call the macro view, treats the courseware objects as black boxes and illustrates some of the communication which must take place between them. It enables us to draw up requirements for the protocols of this inter-object communication. The second view, which we call the micro view, looks inside the courseware objects and illustrates those aspects of their internal structure which enable them to participate in an open courseware system.

The macro view
In its most generic form, the macro view of an open courseware framework can be illustrated as in figure 1.

Figure 1 thumbnail
Figure 1. A generic view of communicating courseware objects

An open courseware framework enables interaction between software entities in order to provide teachers and learners with interactive hypermedia tools for teaching and learning. These software entities may be "courseware resource objects" (CROs) - chunks of courseware of any size. They may equally be "courseware management objects" (CMOs) - applications which are used to exploit the CROs.

Figure 2 thumbnail
Figure 2. A scenario showing two example CMOs (the browse-and-build tool and the librarian tool)

Examples of CMOs would include "browse and build" tools and courseware librarian tools enabling the construction of new "courses" or tutorials, as described in section five and illustrated in figure 2, or they may be "study workbenches" with tools enabling the learner to browse the on-line CROs, communicate via EMail with other learners and create their own documents incorporating resources "captured" while browsing the CROs (and other on-line resources) as indicated in figure 3. These views can all be considered "macro views" of the open courseware framework.

Figure 3 thumbnail
Figure 3. Showing another CMO - the study workbench - and interaction with external tools and resources

The micro view
The micro view of the open courseware framework describes a logical internal structure which all CROs can map to, regardless of their physical internal structures. This logical structure is derived from the HyperCourseware Reference Model. A detailed description of HyperCourseware can be found in [SIVI92]. A brief summary is provided below.

HyperCourseware provides a generic course structure - general enough to accommodate any physical structure which a courseware author can dream up, but sufficiently precise to enable the development of software tools to manipulate it, and to enable unambiguous communication about courseware elements, both amongst people and software entities. This generic course structure can be stated as follows:-

This generic course structure provides:-

A CRO which conforms to the open courseware framework is able to describe its (logical) internal structure in these terms to other courseware objects on request. A CMO must know how to interpret such descriptions and manipulate these structures (e.g., so that it can request a set of Topics from a CRO and do something sensible with the result).


7 Some protocols for the open courseware framework

Analyses of scenarios such as the one outlined in section five has produced a preliminary set of protocols for communication between courseware objects. Most of these result from observing what requests the courseware management objects (CMOs) need to make of courseware resource objects (CROs). Each of these request protocols implies a requirement for the receiving CRO to respond with the appropriate information. The list in figure 4 illustrates some of these protocols in high level form; i.e., it simply outlines what information is being requested without going into detail about the precise format either of the request, or of the information supplied in response to the request.

  • request the type of a courseware object (e.g., a CRO or a CMO)
  • request the type of a CRO (e.g., entire Course, stand alone Topic or stand alone Activity)
  • request the course name of a Course CRO
  • request a Course CRO to launch a Course
  • request if a Course CRO exports a course map
  • request a Course CRO to launch a course map
  • request a set of Topic names from a Course CRO
  • request a Course or Topic CRO to launch a Topic
  • request if a Topic CRO exports a lessons map
  • request a Topic CRO to launch a lessons map
  • request a set of Activity names of a CRO
  • request a CRO to launch an Activity
  • request from a CRO, a set of keywords relating to Activities
  • request from a CRO, a set of keywords relating to Topics
  • request from a CRO, a set of keywords relating to the Course as a whole

figure 4 - Some high level protocols for communicating courseware objects

The list in figure 4 is not exhaustive. We have identified many more protocols at different levels of detail (e.g., requesting a CRO to resume execution following suspension, importing and exporting of navigation histories, and many others). However we hope that this brief list gives the flavour of the kind of protocols required in an open courseware framework.


8 Current state of the research

The generic conceptual structure defined by the HyperCourseware Reference Model is already instantiated as software products on both Macintosh and PC computers. The micro level interface layers and protocols have been successfully implemented and tools have been developed to speed up the building (and maintenance) of CROs. These tools are in use on a number of courseware development projects.

The PC implementation is referred to as "HyperCourseware for the PC". It has proved effective not only as a tool for implementing and managing CROs, but also as a tool for experimenting with the macro level architecture and protocols. It already includes some simple CMOs (e.g., a controls palette which requests and displays Topic lists, manages history mechanisms, requests the launch of Topics, etc.) We have discovered that many of the protocols required at the macro level are actually the same as those already in use at the micro level within HyperCourseware for the PC; e.g., requesting sets of Topics or Activities, launching course maps and lessons maps etc. We are therefore continuing to develop the macro view largely by externalising the relevant HyperCourseware protocols, and only inventing new protocols where necessary.


9 Future research

Although, as stated above, many of the macro level protocols have been established, others have yet to be specified in detail and implemented. Some example issues to be addressed include:-

We see the next stage of research into these issues involving the refining of the protocols and architecture of the macro view, and building more working prototypes of entities interacting with courseware objects at the macro level using the specified protocols. In addition, more work needs to be done to distinguish different types of CMO, and to elicit requirements for protocols to support their use. For example, entities such as maps can be viewed as kinds of CMOs which are dedicated to a particular CRO, whereas a course controls palette is a more generic kind of CMO, but one which most CROs will require at run time. These are both of a different kind to the browse and build tool described earlier, which is strictly an "edit time" entity. If a CRO carries with it a set of "local" CMOs (maps, palettes etc.) then perhaps it should be able to export them on demand to other course objects, e.g., the browse and build tool, so that a lecturer can decide which local CMOs to make available in a courseware session.


10 Summary

We have described what we consider to be severe technical limitations in the approach to the development of the current generation of computer based courseware, and offered a solution to some of these limitations in the form of an open courseware framework. We have described the framework, outlined some of the protocols required to support it, and defined the logical internal structure required of framework comformant courseware resource objects. We have also described HyperCourseware for the PC - a set of tools which embody many of the required features and protocols of open courseware, and which are in use in a number of courseware development projects. Finally, we have outlined some of the issues which we are currently examining as part of our ongoing research.

References

[BUND94] Bundy R., Kilgour A., Norcliffe A., and Siviter P. (1994) CAL for Object Oriented Design, in Proceedings of Hypermedia in Vaasa '94, Vaasa Institute of Technology, pp 98-104
[CTIS93a] The CTISS File, Oxford:CTISS Publications, number 15 (April, 1993).
[CTIS93b] The CTISS File, Oxford:CTISS Publications, number 16 (December, 1993).
[HYPE94] Hypermedia in Vaasa '94, Proceedings of the Conference on Computers and Hypermedia in Engineering Education, Vaasa Institute of Technology (1994).
[LAUR94] Laurillard D., Swift B., & Darby J. (1993), Academics' use of courseware materials: a survey, Association for Learning Technology Journal, Vol. 1, No. 1, pp 4-14.
[LINE94] Linecar P., Siviter D., and Siviter P., (1994). Collaborative Courseware Development for Systems Analysis and Design, Workshop and paper at the XXIX Annual International Conference of the AETT, Napier University, Edinburgh, Kogan Page Ltd.
[RUMB91] Rumbaugh J., Blaha M., Premerlani W., Eddy F. & Lorensen W. (1991). Object-oriented modelling and design, Prentice-Hall International.
[SIVI92] Siviter D., & Brown K., (January 1992). HyperCourseware, Computers and Education, Vol. 18, No. 1-3, pp. 163-170, Pergamon Press plc.
[SIVI94] Siviter D., and Siviter P. (1994). HyperCourseware: Techniques, Tools and Templates for large-scale Courseware Development. In Proceedings of MEDIAactive: Harnessing Multimedia for Higher Education, Liverpool John Moores University.
[SIVI93] Siviter P. (1993). The CAL for Computing Project, in Proceedings of Hypermedia in Vaasa '93, Vaasa Institute of Technology, pp.261-265.
[TRAI92] Trainor R. (1992). Computers, arts based teaching and rising student numbers. In The CTISS File, University of Oxford: CTISS Publications, number 13, pp 3-6.

ACE publications : Published 1994 - Copyright the authors. All rights reserved.