Authors: Jacques Levin, Ph.D., James Barker, M.S., Mark Eaton, M.S. & Dale Perrin, M.S.
Since 1983, Nova Southeastern University (NSU) has used computer-mediated communications (CMC) for its distance learning environment. Like many institutions engaged in this form of distance learning, the CMC model has been one of a centralized structure where NSU provides a hub for the network. The centralized model is much like the wheel of a bicycle with the hub of the wheel representing the university computing systems and the rim representing the student computing environment. Connectivity is provided by telecommunications (the spokes of the wheel) with control of the network residing at the hub. In this centralized model, all is dependent on the university system. The problem with this network design is the university hardware and software become the focal point of the learning environment rather then the faculty or students where the collaboration of research and learning actually takes place.
Under the current operational environment at NSU, faculty and students connect to the network through an AT&T service using their Mac, PC, or Unix based machines, which then become effectively Unix remote terminals off of the NSU distance learning network. The system has been very effective for text based communications and has facilitated some creative applications such as the Electronic Classroom Room (ECR), where faculty can hold virtual classrooms with students in attendance; Electronic Library (EL) that can be used for research and submitting requests for documentation; and regular email for administrative and academic communications.
In 1995, NSU formed a research group of post graduate students to study the application of Internet/Web tools for bringing multimedia to this learning environment. Scigliano, Levin, and Horne (1995) reported the effort to improve interaction on the NSU distance learning network through some of these new Web tools (e.g. Hypertext Markup Language or HTML). Faculty and students are now authoring with HTML extensively thus developing a non-linear knowledge base that resides on the NSU server and which facilitates collaborative learning and research. But the very nature of such a collaborative learning environment is that it is not centralized and we see this in the way the Internet is evolving and structured today. Consequently, the multimedia research group formed a special project team earlier this year to investigate the possibility of distributing the learning environment under a concept known as "server-at-home".
This paper reviews the problems and vision that led to the establishment of the project team and how some preliminary results from them spawned an environment of computer-supported cooperative work (CSCW) which is being applied in other related research team efforts under the multimedia research group at NSU.
Much of the material developed by faculty and students is published via the university on-line system. This centralized model requires faculty and students to adhere to the requirements established within the university computing environment. Users must compete with every other user for central processing time, as well as mass storage space. They must also restrict themselves to a single operating system and use only tools approved by managers of the computing resources. The limitations of the centralized model are most evident when many users are accessing one server simultaneously. In the current environment, all work produced by students must be transmitted to faculty via the university hub and this can cause a number of problems. Students work in a variety of operating systems off-line to include MS-DOS, Microsoft Windows and MacOS as well as less common operating systems such as NeXTSTEP, BSD, and Solaris for PCs. Few of these hardware and software combinations can directly interface with the university computing environment so students must convert their work from these various environments to a format compatible with the university operating system. Frequently this conversion results in less then perfect representations of the products that are ultimately to be presented to faculty for grading.
There are also hardware related problems of concern. If the university's computer system, which literally mediates the network, fails for any reason, all faculty and students are adversely effected. Users are affected by connectivity failures at the university end or when the server is overwhelmed with traffic to the point of non-responsiveness. Viruses, hackers, and any other type of computing hazard have equally deleterious effects on all users in the centralized model.
Forcing faculty and students to comply with a centrally controlled operating environment also affects the collaborative learning and research environment. Basic HTML publishing can be easily accomplished in the current environment but advanced HTML work (e.g. HTML forms requiring common gateway interface [CGI] support) require extensive assistance from system administrators whose frame of reference is completely different than those of faculty and students. They must be concerned with the impact of CGI scripts on the welfare of the entire system. The progress of research projects are negatively affected by this bureaucracy especially in a distance learning environment where staff, faculty and students are geographically separated.
It is clear that a better model is needed if it can be economically and reliably accomplished. The vision for this new model calls for a completely distributed environment where each participant in the model is both server and client in the client/server paradigm with connectivity through the World Wide Web (WWW). While we are on the threshold of implementing this vision, it stills eludes us for a number of reasons.
The first challenge to implementation of the distributed model is the technical difficulties associated with the installation and configuration of the operating system (Welsh, 1995) and utilities necessary to operate a Web server (Callahan and Longyear, 1995). The solution must be easy enough for a distance education student to implement, yet substantial enough to support the requirements of the model.
A second major obstacle concerns acquisition of Internet Protocol (IP) addresses. The current 32 bit addressing schema, referred to as Internet Protocol version 4 (IPv4), is very nearly out of addresses so providing server capability to each participant is extremely limited since they must compete with potential server acquisitions world wide (Hinden and Deering, 1995).
The next generation of Internet Protocol (IPv6) is intended to overcome this problem and will support Internet traffic for many years into the future by providing enhancements over the capabilities of the existing IPv4 service (Hinden and Deering, 1995). IPv6 addresses are 128 bit node identifiers composed of eight 16 bit hexadecimal values separated by colons.
The critical points of the emerging IPv6 standard as they relate to our vision for this new distance education model; are two. First, the standard includes a migration plan which defines a method for the transition and mapping of IPv4 addresses. NSU will eventually have to comply with the IPv6 standard, however until that migration occurs the IPv6 standard will support current IPv4 addressing schema. Second, based on the proposed plan for the implementation of the IPv6 standard, there is nothing in the current plan which would preclude the successful execution of the distributed distance education model on a smaller scale.
As a result, the project team is prototyping an interim phase for the vision where participants within a common domain (in this case, the project team) have access to the Internet as clients and at least one of the participants function as both server and client in the client/server paradigm. This interim plan at least provides the primary benefit of having server capability so that researchers can control and exercise collaborative activities. The interim plan allows for more than one participant to have server capability and more are being added as research members are establishing their own server-at-home capability. Establishing server capability requires two elements.
First, any server must run a WWW server program (Hart and Rosenberg, 1995). The most popular WWW server for UNIX systems, for example, is a daemon named httpd developed by the National Center for Supercomputing Applications (NCSA). The httpd is freely available on numerous web sites. Each WWW server program (e.g. the httpd daemon) has an associated group of configuration files, which allow the system administrator to fine tune the server.
Second, each server site must provide a home page program (Hart and Rosenberg, 1995). This serves as the starting point for clients (faculty and students) to explore the server. The server's home page would most likely contain hypertext documents (or possibly Java applets) that point to data and services available on the server or possibly services available on other servers. This is a well understood process and is already beginning to be implemented under the existing centralized distance education prototype.
On the client side of the model the only software necessary is a web browser program (Hart and Rosenberg, 1995). Faculty or students can use any browser they wish. The choice of browser is purely a personal matter of style. Among the most popular are Netscape and NCSA's Mosaic, as well as several proprietary browsers. Netscape and NCSA's Mosaic, for education and research, are freely available on numerous web sites.
Nova Southeastern University has facilitated the study of the model by establishing a Point to Point Protocol (PPP) line for use by project team members. A PPP line allows transmission of IP datagrams over a serial line connection (Kirch, 1995). NSU has established PPP as its standard for a TCP/IP dial-up connection, therefore software necessary to establish a PPP connection is also an essential to the model. Each project team member has Internet access capability and several are establishing server capability on their home machines within various operating systems like LINUX and Windows NT. The prototype effort is being managed through six phases: 1.) Setting up prototype servers. 2.) Installing/configuring a conferencing application on at least one server within the common domain (in this case, the project team). 3.) Preparing a "Welcome" page on the primary server. 4.) Setting up prototype clients. 5.) Monitoring clients connected to the server 6.) Conducting client\server connection tests.
One server-at-home was used to setup conferencing capability for researchers to use in exchanging notes and collaborating on their studies. This was deemed essential since the new model of all participants having server capability is still some distance in the future. Conferencing capability within the research domain is essential to foster the collaborative work environment and provide distributed project management of the implementation of the new distance learning model.
The conferencing package selected was WebNotes from OSTECHnologies Corporation. Perhaps the most compelling function of WebNotes is that any user can create and conduct conferences directly from Web browsers. WebNotes simultaneously serves native format, standard Web documents, and images. This makes WebNotes a practical, useful, and reliable tool for managing projects across the Internet. When conference participants log in through their Web Browsers, a series of buttons allows access to view, reply, search for notes and forums, gain on-line help, and build server management skill. Installation of the software was fairly straightforward. The use of a this interim phase (involving server capability and WebNotes conferencing capability among researcher's) resulted in a sensible, open architecture that consists of interacting home-based microcomputer systems working across the Internet, with emphasis on a concentric approach to implementing technical and management processes for performing on-line graduate research. Further, this framework captures information concerning the establishment of procedures for planning, monitoring, controlling, reporting and assessment in the emerging field CSCW. The framework supports the use of established project management techniques in CSCW and increases knowledge about how to improve and manage inter-project communications.
The "concentric, cooperation and control" or C3 Project Control System increases the availability of computer power to individuals and places tools of project management within the reach of every user. Project team members with Internet connections and WWW browsers, regardless of the barrier of time and distance, can share information, build up a knowledge base or information base, and maintain close contact. Having the interim prototype phase at hand extends the principle, use and application of distributed project management.
With the rapid growth of knowledge, is the growing need for project teams to add to and distribute knowledge where and when it is needed. The new model framework addresses the issue of inter-project communication, that is a critical issue for the NSU multimedia research group which consists of several projects teams. In conducting these multiple simultaneous on-line research sub-projects there is the aspect of creating a "total system" design for the future NSU distance learning environment, and relating several underlying interwoven bodies of knowledge. The framework makes it certain that each project team can profit by sharing knowledge and insight developed by the other project teams.
We've shown here that the current CMC distance learning environment at NSU has been effective but limited in collaborative research activities. A new model was proposed where each participant is both server and client using Web tools to communicate and build research knowledge bases. Establishing server capability is quite challenging especially to distance learners who lack proficient computer skills, plus the current IP addressing scheme prevents the rapid growth of servers that would be needed to support the model. These problems will eventually be delimited, but until then, the NSU server-at-home project team is prototyping capabilities for research students to establish their own server capability and build a collaborative learning environment among the entire multimedia research group. The prototype has resulted in a computer-supported cooperative work environment which is beneficial to related projects in establishing the future distance learning environment at Nova Southeastern University.