The approach taken for this chapter is to first define the unifying term distributed open learning, to be used in further discussion. The requirements for distributed open learning environments will then be identified. A review of existing environments will reveal how well they meet those requirements. Scope and objectives of current standardisation activities will be examined. As a conclusion of this chapter, directions of DOLE standardisation will be discussed.
The process of learning in the context of education has been well defined. Traditionally, learning process was done at special places, such as classrooms or schools. Distance learning was introduced to extend the learning process outside these places. The main problem with distance learning was that it still has the same paradigm as traditional learning: instructor-centred. This paradigm has the view that learning is one-way, from the educator to the learner. In contrast, the newer concept of learner-centred paradigm advocates that the learner explores the knowledge on their own, and set their own pace. The educator only facilitates during the process. When combined with co-operative or collaborative learning, where the learner forms small teams, there are added benefits such as (Marjanovic, 1996):
If distance learning is conducted with the learner-centred paradigm, the learning process can happen anytime and anywhere, giving rise to the concept of open learning. Although many papers such as (Marjanovic, 1996), (Chou & Sun, 1996), and (Abdel-Wahab, 1996) still use the term distance learning, we believe that this term carries too many classical connotations with it. Indeed, some distance learning institutions now call themselves providers of distributed learning, most probably to appear modern and forward-looking. However, the term distributed learning is also used in other context, such as (Lotus, 1996), who define it as a "technology-enabled, learning-team based focused education, facilitated by a content expert, and delivered anytime and anywhere."
Since there does not seem to be a consensus on the term used (Sener, 1997) and to avoid confusion, we will define a unifying term: distributed open learning. This will include collaborative learner-centred paradigms, anytime - anywhere delivery, and distance learning itself. It also our intention that distributed open learning includes other important concepts such as lifelong learning (Slonim, 1996), sometimes also called continuing education, and just-in-time learning (Schoening & Wheeler, 1997).
We will further define computer-based learning environments that support distributed open learning to be distributed open learning environments (DOLE). For a perspective on these environments, the requirements need to be identified first. There are two types of requirement: technology and socio-economic requirements.
Papers such as (Lotus, 1996), (Abdel-Wahab, 1996), and (Slonim, 1996) has listed many technological requirements for distributed open learning environments. The common requirements listed are:
Alongside technological requirements, (Slonim, 1996) and (Arvan, 1997) also identified social and economic requirements, such as ownership or copyrights, cost, security, and privacy. The issue of cost is also discussed in detail in (Garson, 1996) and (Turoff, 1997). There is also the issue of "education standards", the measure of quality of education in one institution as compared to another (Vroeijenstijn, 1994). These socio-economic requirements of distributed open learning are out of the scope of this project and will not be discussed further.
There are two basic model for DOLE:
The classification of DOLE outlined here is based on two layers. The lower layer is the underlying communication layer, called communication and networking infrastructure. The upper layer is the application, further divided into Internet-applications-based and groupware-based. Each layer may consist of the synchronous and/or asynchronous model.
The examples of DOLE reviewed here is only a representation of existing environments. For an exhaustive look at the use of computer networks for learning, see (Harasim et. al., 1995). Unfortunately, it is biased toward the more classical, proprietary networks, instead of the current trend of using open networks such as the Internet.
This section will review examples of work involving communication and networking infrastructure for distributed open learning. Most of the literature in this area actually has titles and objectives that claims to design a fully functional learning environment. On deeper inspection, only communications or networking aspects of the environments are considered or emphasised. Nevertheless, communication and networking infrastructure is an important part of distributed open learning.
Agent technology is currently being proposed for doing the routine works during collaborative processes, or even assisting learners. For example, Grammar Collaborative Intelligent Learning Environment (GRACILE), as described in (Ayala, 1996) is an "environment where intelligent agents communicate and cooperate by exchanging messages and providing the learners with access to a larger body of knowledge, finding assistance possibilities and promoting the effective collaboration in the learning group." Another example is Multi-Agent Processing Environment (MAPE) (Shi, 1996), which is designed for computer-supported co-operative work (CSCW), but is applicable to learning environments. Agent technology originates from the field of artificial intelligence, and its application in distributed open learning is still at an early stage.
In these section examples of systems that extends classroom using computer networks are reviewed. Interactive Remote Instruction (IRI) as described in (Abdel-Wahab, 1996a), and DooRae-EDUS (Park, 1996), aims to provide facilities for multimedia-supported interactions (video, audio, data sharing), presentation tools, surveys, evaluation, homework, multimedia note taking, recording and playback of classes. These environments are certainly useful, but usually require a local area network (LAN) with high data rate. This means they are not suitable for large scale (either in terms of number of learners or distance) distributed open learning. An exception is (Tzeng, 1996) who described a course delivery mechanism designed for public broadband networks. However, the mechanism only deals with multimedia materials, with limited interactivity and no collaboration involved.
The environments that are grouped into this category have more consideration of learning aspects, instead of concentrating solely on communication or networking technology. For example Learning Environment (LE) (Fretwell-Downing, 1996) is based on a repository of information, enabling learners to quickly locate and use course materials. However, LE does not support collaboration effectively. Environments such as Hyper-G (Skillicorn, 1996), and Co-operative Remotely Accessible Learning (CORAL) (Chou, 1996) provides hypermedia-based tutorial, with facilities for navigating, annotating and discussing the material collaboratively. CORAL also supports synchronous virtual classroom activities like audio and video conferencing, shared whiteboard, and chatting. Environments such as these may provide a starting point for generalised models of distributed open learning.
The term Internet applications-based is used because the usual term Internet-based can be misleading. Customised environments as discussed in section Customised environments, or groupware-based (see section Groupware-based) can (and most do) use the Internet as the underlying network fabric. When the literature says Internet-based they usually mean Internet applications such as World Wide Web (WWW), chat, electronic mail, Mbone, etc. The main underlying technology used is usually the WWW. These WWW-based learning environment seems to be current trend, as shown by the large numbers of current literature presenting it. The WWW is then either used on its own, for example (Dumont, 1996), or enhanced with other applications to improve collaboration, interactivity, and management. For the purpose of this project, we will focus on the latter type, as it is more closely related with the concept of distributed open learning.
Internet applications-based learning environment such as WebCT (Goldberg, 1997), Virtual-U (see Virtual-U homepage), and Internet Learning Environment (ILE) (Lockledge, 1996) usually rely on scripts or programs that are run on the WWW server. Course materials are presented on WWW pages, and the environments provide support for discussion, either asynchronous (by posting messages to one or more "conferences") or synchronous ("chat"). Learners can access the environments by using any WWW browsers, and connecting via the Internet or campus network.
A good definition of groupware is found in the classic paper (Ellis, 1991): "computer-based systems that support groups of people engaged in a common task (or goal) and that provide an interface to a shared environment." The literature on groupware concentrates mostly on using it to support business tasks such as office automation, meetings, and decision support systems. So there are less current research literature of using groupware to create a learning environment, compared to the WWW-based learning environments. It seems that the proponents of distributed open learning are touting WWW as "the" solution. One notable exception is (Marjanovic, 1996), who advocates the use of groupware technology instead of WWW, arguing that "groupware is designed to support interaction and collaborative problem solving" and "offers rich media types, better security, object link maintenance, easier development of applications and full interactivity". The issue of using either groupware or WWW will be elaborated more on section Combining groupware and WWW.
Currently it seems that the only example worth mentioning in this category is Lotus LearningSpace (Lotus, 1996), which is based on Lotus Notes, an established groupware product. The LearningSpace application is composed of interconnected modules, each of which is a Lotus Notes database:
Learners receive the
The argument whether to use groupware or WWW for collaboration has surfaced many times, for example in (Borysowich, 1996) and (O'Connell, 1995), although usually not in the context of distributed open learning. However, some of the results of the comparisons are relevant, as it highlights the strength and weaknesses of each technology. One of the important issue is the support for occasionally-connected learner. Groupware technology usually supports some kind of replication technology, where clients can synchronise its database contents by connecting to the server. When the clients disconnect, they will have the same content as the server, so the learner can use it anywhere without being connected.
WWW-based environments require the learner to be well-connected. On the other hand, the fact that a learner only requires a WWW browser, now widely available on many platforms, to interface with WWW-based environments is appealing. So we believe that the question should not be whether to choose WWW or groupware, but rather how can we combine WWW and groupware.
All the environments discussed meets most of the technology requirements stated on section Technology Requirements, to a varying degree:
However, one particular requirement that was not in the design of those environments is standards-based. Internet-applications-based environments, particularly WWW-based received more attention than other technology because it is seen as platform independent (browsers available on all platforms), and use a standard language (Hypertext Markup Language, HTML) for description of course material.
Unfortunately, this is where the standardisation stop. The WWW was originally designed only for information dissemination through distributed, hypertext-like documents. To add seamless multimedia capabilities to WWW pages, several approaches are now used, such as audio or video streaming, using Java language, or browser extensions. These approaches are not standardised, and not as widely available as browsers.
Furthermore, to meet collaborative requirements, more non standard extensions are made to the WWW server, making course developed on one WWW-based environment not usable on other environment. So even starting with what looked as a universal platform does not guarantee standardisation.
In order for DOLE to be widely used, and to eliminate interoperability problems, some organisations are sponsoring or conducting activities to define standards. The standards activities outlined here are technical or computer standards, not educational standards. Technical standards address issues mainly related to software systems, while educational standards address issues such as curriculum contents and quality of education (see section Socio-economic Requirements).
This group is sponsored by the Institute for Electrical and Electronics Engineers (IEEE) Computer Society Standards Activity Board, intending to develop a series of standards for computer-based learning (Schoening, 1997). It will try to develop standards, guidelines, and recommended practices for the area of computer-based learning, with the goal of enabling tools, courseware, information, and services to be provided on a component basis. P1484 currently has nine working and study groups, with more groups expected:
RM-ODP aims to achieve (Linington, 1995) and (ISO/IEC, 1995):
The reference model provides a "big picture" that organises the pieces of an ODP system into a coherent whole. It does not try to standardise the components of the system nor to unnecessarily influence the choice of technology. The RM-ODP standard is known as both ISO International Standard 10746 and ITU-T X.900 Series of Recommendations and will consist of four parts:
Part 1 contains a motivational overview of ODP and explains the key concepts of the RM-ODP architecture. Part 2 gives precise definitions of the concepts required to specify distributed processing systems. Part 3 prescribes a framework of concepts, structures, rules, and functions required for open distributed processing. Part 4 describes how the modelling concepts of Part 2 can be represented in a number of formal description techniques.
Although originally designed for distributed processing systems, the RM-ODP standard is actually generic or abstract enough to describe any kind of application. This means that it may be possible to use RM-ODP to define a framework for DOLE. Using RM-ODP in such a way will eliminate the need to create another reference standard.
OILS (BESA, 1996) is a UK initiative, involving more than 30 companies and educationalists. It is a set of agreed rules governing the structure and operation of the underlying system of monitoring and recording, including methods for setting up workplans, gathering and assessing user responses and the format for storing data in disk files. See also OILS Standard) for more information.
CEdMA (see CEdMA homepage) is a US-based association for professional association for individuals who manage companies that produce computer-based education systems. CEdMA has an initiative titled Learning Architecture and Learning API Task Force. This initiative is a component-based approach for standardising computer-based learning environments, focusing on independent Learning Objects.
IMS (see IMS Project) is a project by the National Learning Infrastructure Initiative (NLII), which is in turn sponsored by Educom. Both NLII and Educom are US-based organisations. IMS will create standards that enable WWW-based instructional objects to interoperate, with common mechanisms for organisation and retrieval.
This committee (see AICC site) has published guidelines and recommendation to enable the interoperability of computer-managed instruction (CMI) systems. This means that a given CMI system can manage computer-based tutorial (CBT) lessons from different origins. Also, a given CBT lesson can exchange data with different CMI systems.
There is an effort by major software companies to produce component interoperability, such as ActiveX from Microsoft and Java from Sun Microsystems (Roschell, 1996). Currently these architectures are still focusing on language issues, and not yet considering higher-order application components. Work is also being done by the software engineering community to provide reusable software architectures (Meyer, 1997, chap. 4).
The guidelines and recommendation of AICC are very basic and specific, and AICC only has influence on the aviation industry. Component software architecture is currently focusing on overcoming language and platform barriers, so they are still at a very low level of software systems. OILS have a very practical approach to standardisation, basically focusing on file formats, and expand from there. IMS has a wider scope, but still it only focuses on WWW-based environments. This means IMS is too technology-dependent and only affects a subset of DOLE. Nevertheless, all the activities mentioned here will be useful as an input to wider standards activities.
Both RM-ODP and CEdMA have a potentially large impact on DOLE. RM-ODP is an established international standard, specifically designed to accommodate instances of application domain frameworks. Currently no effort has been made to use RM-ODP to describe DOLE. CEdMA has a wider scope than IMS or OILS, and is on the right track to creating a framework for distributed open learning environments. Like OILS, CEdMA follows a practical and therefore limiting approach to standardisation. So RM-ODP and CEdMA are classified as having medium impact on DOLE.
IEEE P1484 is probably the most ambitious activities, with an aim to create a standard for every imaginable aspect of computer-based learning. Backed by IEEE, widely known for its role on other technical standards, it has the chance to have the largest impact on distributed open learning, compared with other activities. However, it has been noted in the P1484 discussion groups that its effort originates from academic research and currently lacks support from the commercial world.
The standardisation of DOLE are driven from two directions: from the bottom-up and top-down. The bottom-up direction is standardisation of the communication and networking infrastructures. The top-down direction is standardisation of the software systems used for learning. Since one of the goal of DOLE is universal access, it is also important that both directions of standardisation considers scalability issues.
The standards for computer networks are largely in place. Standards (both
de facto and ratified international standards) for local area networks
(LAN), metropolitan are networks (MAN), and wide area networks (WAN) exist.
Protocols used on those networks, although numerous, has boiled down to one
or two major protocol. One often cited example of an accessible and very
large scale network is the Internet and its corresponding de facto
protocol, TCP/IP
Standards in communication activities such as synchronous audio and video conferencing are emerging (Ginsburg, 1996, chap. 2). Besides the usual international standards body such as International Standards Organisation (ISO) and International Telecommunication Union (ITU), other organisations such as Asynchronous Transfer Mode (ATM) Forum and Digital Audio Video Council (DAVIC) are also trying to achieve standards in this area.
Since most of the networking standards are established, a large commercial networking market is in place. There already numerous companies producing communication infrastructure for DOLE, in the form of conferencing systems, groupware, and hardware. This market is a powerful driving force towards better networking standards.
Standards for software applications that drives DOLE is still lacking. As discussed in section Component-based software architecture, there is still no agreement on how to break a software into components, and make them interoperable. Terms like components, frameworks, and patterns are being used to describe research in this area (Johnson, 1997).
The main driving force in the top-down standardisation of DOLE is likely to be the academic research community, as shown by the ambitious goals of IEEE P1484 (described in section IEEE P1484. Other areas such as research in intelligent tutoring systems (ITS) aims for "smarter" DOLE. ITS aspects that needs to be standardised include learner models and task definitions.