Last update at : Tue May 9 7:38:55 1995

Cyberspace Learning Community Hawaii

Analyzing Linkage Structure in a Course-Integrated

Virtual Learning Community on the World wide Web

Home Page of Leon James
Go to the bottom
Home Page of Kevin Bogan
See a related paper by Leon James
See the students' reports

Are you a visitor? I would appreciate your reactions or comments. How did you get here? What did you see?
Leave a massage for Leon James?

Leon James, Professor of Psychology
Kevin Bogan, Department of Psychology
University of Hawaii

Table of Contents

Hypertext as a Cultural Resource
Mining Hypertext as a Learning Resource
Creating the Generational Virtual Superdocument
Focus on Learning Skills
The Weekly Assignments
The Expanding Superdocument
The Virtual Size of a Hypertext Superdocument
Structural Properties of Hypertext Links in a Cyberspace Learning Community
Research Issues
Natural vs. Artivicial Superdocuments


College students are participating every semester in a regular psychology seminar in which their weekly homework is submitted as HTML documents linked to their individual Home Pages which are located in a subdirectory attached to the instructor's Home Pag e on the World Wide Web. Students read each other's reports and create HREF links between their comments and the target paragraph in the other student's Web document. Each semester the files are archived and linked to the current students' Home Pages. Students read and react to all the prior generation files, creating new links between themselves and prior generations. Anachronistic links are also created by the instructor between prior generations to subsequent ones. The result is the natural growth of a generational cybercommunity embodied in the form of a hypertext superdocument that, in theory, can develop endlessly, creating an evolving educational cyberspace nook on the World Wide Web that can be revisited, again and again. Though hypertext te chnology was designed to facilitate reading and presenting complex documents, it appears from this on-going experiment that the application of hypertext links can be extended within an educational context to create cybersocialization forces that aid in th e learning of communicative skills, develop critical thinking, and motivate the acquisition of lifelong information literacy.


The authoring of documents in an educational environment traditionally has relied upon manual references to link one work or set of statements to another. Dictionaries and encyclopedias are simple examples of a linear usage [1] The works of Aristotle of ten refer to other works and nested clarifications and annotations in the Talmud show a more intricate form of reference [1]. In some instances, within-text references frequently are indicated with a reference to a former or latter statement. With the a dvent of computers, machine-supported links became possible.

As early as the 1960s, when Theodor Nelson coined the term, "hypertext," programmers began developing material that used machine-supported links [1], [2]. Hypertext can be described as textual material, from passages to keywords and headings to whole doc uments, whose parts are connected by links allowing paths that can be traversed in a non-linear manner [1], [2], [3]. Hypermedia has the additional ability to link to other media such as images, sound clips, video, graphics, executable computer programs, and more [3]. At its basic level, links may be described as either relational or organizational interconnections of information [1], [4]. Relational links are typically defined by either the author or the user to suit the particular use in mind. In th e Issue-Based Information Systems (IBIS) devised by Horst Rittel and his students [1], eight types of relational links are used: those links that allow the author to respond to an issue, those that pose a question to an issue, those that support, those t hat object, those that refer, those that replace, those that generalize, and those that specialize [1]. Organizational links structure the links in a controlled fashion, such as hierarchical (tree), alphabetical, or functional (skipping to a particular p oint either forward to the bottom or back to the top) [3]. Conklin [1] details the use of 20 different systems for hypertext and hypermedia. Most were designed as on-line documentation; writing, research, and design aids; and educational aids. Of those designed for education and research in a university setting, Intermedia [2] was developed at Brown University's Institute for Research in Information and Scholarship (IRIS) as an aid for both readers and authors.

Intermedia [2] uses an object-oriented, unix-based system of five integrated applications. Most of the applications were taken from the Apple Macintosh suite of programs to provide editors for bit-mapped images, text, and timeline, along with viewers for three-dimensional objects and scanned images. Consistency of use is maintained by using the same operations throughout the different programs. For example, double-clicking to open files and programs was extended to double-clicking on links to open them . This consistency of use made the system relatively quick to learn. As of 1988, Intermedia had been used to teach an English class and a cell biology class. The English class had eight users qualified as authors with an additional 80 students using th e system primarily for browsing. Only a few students attempted writing their own documents. The major problem experienced in using Intermedia was having to work in a heavily layered environment: Cadmus Computing's CadMac toolbox; Inheritance C, an obje ct -oriented programming language; Apple's MacApp which allows creation of standard applications in the Macintosh manner; and a number of other applications designed by the Intermedia team of developers. As an educational tool, it performs well providing for multiple users (authors may work on a document simultaneously) and multiple documents may be open at any one time. Multiple windows on a screen provide graphic display of the links. Intermedia has proved to be an excellent tool for instructors to c reatively pass information to students. However, the path of knowledge remains linear leading from instructor to students. There is no kindling of ideas from student to student, or from one generation to another. Only recently has hypertext in the clas sroom shown the potential as a resource for linking together various areas of thought, knowledge, information, and beliefs.

This experiment has demonstrated the adaptability of hypertext technology to an undergraduate college setting that provides course-integrated use of the World Wide Web. It has shown that first time users can independently acquire the skills involved in c reating a Home Page and in using file transfer software to upload their assignment reports. The procedures developed for the course were intended to facilitate the creation of an enduring, generational learning community whose virtual existence was in cy berspace in the form of a hypertext superdocument created, used, and maintained by each succeeding generation of students in the course. The interlinked student reports were specific, topic related connections perceived by students as they read each othe r's reports throughout the semester. Close observation of the generational, cumulative hypertext superdocument promises to yield data on the structure and development of educational cybercommunities. The analysis of links shows that they fall into a few functional types indicating that interpersonal hypertext linking has social significance within the cybercommunity.

When a student decides to create a link from self to other, a connection is proposed and established on the basis of the semantic content of the passages being linked. Within the set-up for this experiment, students engaged in interlinking each other as a weekly assignment, though each student chose where to make the links. Interlinking thus became a regular and normal social transaction. The future study of hypertext interlinking as a social transaction within cybercommunities, will help us understand the global Internet phenomenon and its potential significance for society and social evolution.

Although the creation of electronic hypertext is a recent event, its cultural antecedents are easy to trace since all scholarship and knowledge in the past has also proceeded through a hypertext architecture, though at a more primitive level of technology. The recognition that human knowledge is created through the connection of ideas has been expounded in detail by Aristotle, Swedenborg, Locke, Hume, John Dewey, Freud, Skinner and many other psychologists and educators. Scholarly writing and research is traditionally distinguished from popular and other writings in terms of the required presence of bibliographic citations, whose function it is to interlink similar work for the purpose of comparison, exposition, and objectivity.

The interlinking of information has received a central focus in on-line catalogs [5} through keyword searching, a procedure that links entries in fields by means of common tags (or anchors). Cross-referencing in subject heading lists (or 'see also' connections) are linkage devices that bring together text from independent sources. Subject bibliographies provide lists on related topics that allow access to independently published sources. In the print medium, links have been created within documents using various approaches, including quotations, citations, reproductions, and footnotes. All of these devices share the function of linking some text or reproduction to some other text or reproduction. When previously unlinked ideas are linked through hypertext knew knowledge comes into existence, new products and resources, possibly a discovery or invention.


Since the recent advent of electronically linked multimedia hypertext on the World Wide Web, the creation of knowledge through interlinking ideas or information has taken on new significance for culture, education, and commerce. It is now instructionally feasible as a routine activity to build learning communities in an academic setting. Course-integrated use of the Internet opens up new educational possibilities. Students can now produce Web documents with hotlinks to other students and to whatever the instructor's assignments direct them. The electronic hypertext environment, accessed through the information superhighway, provides the medium in which students can practice authorship, critical analysis, and social development through community integration.

A critical factor in the success of a cyberspace learning community is the ability of the instructor to create community-building forces within the class. A group of people who are given a joint task to accomplish can find motivation in group solidarity or competitiveness. Because few can accomplish complex tasks on their own, a socially organized framework of mutual assistance needs to be put in place so that no one is left behind in failure and embarrassment. A visible communal product to which all contribute, and with which all can identify, needs to be developed and perceived as the outcome of their labor and effort. In short, the cybercommunity created within an academic context needs to be real, believable, and authentic. To the extent that it has these properties, to that extent the virtual learning community [6]will be effective. The educational experiment described here will be examined to see how these community-building forces were built up to maintain effective cybersocialization practices, and what instructional management mechanisms need to be maintained in order to foster the growth of a cyberspace learning community.


The Set-up
Students enrolled in a psychology seminar discover at their first class meeting that all course assignments are to be submitted online through the creation of Web documents. Only rarely can a student be found who has ever made use of the Internet or performed related activities such as file transfer or electronic mail. Almost all are at the level of feeling comfortable doing their homework on a word processor and doing minimal file management on a desktop computer. The technical tasks expected of them that are briefly described by the instructor seem unrealistic and unreachable to most of the students: uploading, ftp, e-mail, telnet, HTML, emacs editor, UNIX file management. They are given a login account on the Web server of the College and the address of several campus computer labs from which they can have access. Only about ten percent of the students use a dial-in modem.

Students receive no formal instruction in accessing the Internet or creating their Web documents. All Internet instruction is handled through online instructions and in the regularly scheduled class meetings. Typically, about half of the two-hour class is devoted to handling Internet problems and the other half is made up of activities relating to course content. The face-to-face class discussions serve as a forum for venting frustrations and a source of mutual assistance. Weekly assignments published on-line on the instructor's Home Page force a certain pace that students must maintain or else fall behind, which carries a point penalty towards the course grade. The pressure students feel during the first four weeks is reportedly intense. This anxiety is expressed orally in class, on e-mail to the instructor, and in comments students make in their written reports.

Two specific instructional techniques were used to assist students in overcoming technophobia, resistance, and the impulse to quit. One method was to encourage students to help one another. For example, a discouraged student in panic would be teamed up with another student who was more skilled and confident, and they would spend one or more sessions together in the lab. Another student from a prior semester volunteered to be at the lab during certain announced hours so as to provide one-on-one help. The second method was to have students give oral presentations on the content of student self-witnessing reports from prior semesters. This feature was called "studying the generational curriculum." Reading, presenting and discussing the self-witnessing reports of former students facilitated their understanding of what is expected of them and allowed them to vent their emotions since these reports are full of desperate descriptions of being in panic. Best of all, the generational student reports are almost invariably up beat; they end in success and enthusiasm, thus, a complete turnabout. The generational self-witnessing reports are the proof that (a) they are not alone in feeling the desperation of infoshock and (b) they have overcome and reached success, so therefore they can too.

Focus on Learning Skills

The skills students acquire through this course-integrated cybercommunity fall into three categories: (a) educational networking skills; (b) information literacy skills; and (c) scientific or scholarly skills. In terms of networking skills, students achieve the following significant steps by midterm: (1) they stop feeling panicky or depressed and start feeling enthusiastic and self-confident; (2) they learn to use e-mail and a UNIX editor (either emacs or Pico); (3) they learn to upload their assignments from their word processing diskette and to translate their text into HTML format; (4) they learn to use Netscape and Lynx and to search the Internet through gopher indexes, Veronica searches, Web directories, and WebCrawler searches; (5) they create a Home Page on the World Wide Web giving access to their weekly assignments which by now averages about 40 screens with 50 hotlinks and a hundred anchors.

In terms of information literacy skills, students practice the following throughout the semester: (a) translating their problems into written form so that they can receive a solution in e-mail form; (b) following complex online instructions and executing the sequential steps required; (c) finding information in a hypertext environment; (d) strengthening one's courage and resolve in becoming a perpetual but successful novice in the ever-changing kaleidoscope of information systems. In terms of scientific and scholarly skills, students are engaged in the following: (a) learning to write in public and for an interactive purpose; (b) learning to perform critical analyses of other people's writings; (c) maintaining an intellectual presence or voice within a generational learning community through citizenship activities in cybercommunity, e.g., maintaining a Home Page on the World Wide Web and acting as host to cybernaut browsers both inside and outside the learning community; (d) making individual contributions to the building of a virtual superdocument and thus meeting the responsibility of being a published author; (e) learning to express one's intellectual position on pre-assigned topics.

The Weekly Assignments

The detailed weekly pacing of student work is essential to maintain instructional control and organization of materials. Homework for the first few weeks directed students to engage in acquiring basic skills such as logging on, using communications software, sending e-mail, and navigating the information superhighway through Lynx, Netscape, and Gopher. The second phase directed students to upload their reports and to use the online editor to insert HREF links to other students' reports. A major effort is required to help students gain an understanding of relative and absolute URL paths, and this topic comes up at every class meeting until midterm and beyond. The interlinking of student reports is made part of the assignment every week. Students access each other's Web documents through the instructor's Home Page and make a link from one paragraph in their report to a paragraph in someone else's report. Students are free to choose who they link to and what about. Mutual two-way links are strongly encouraged so that students get into the habit of searching other people's documents, and when they find that someone has made a link to their report, they then go to that anchored paragraph in their own report and create a link to the other student's anchored paragraph.

As depicted in Fig.1, the instructor's weekly homework assignment document occupies a central role in the cybersocialization effort by which students become functioning members of the cybercommunity. This document serves as the medium for announcements, instructions, and advice. It introduces new content, re-interprets old content, and responds to ongoing concerns and themes in the cybercommunity. The instructor's document contains encouragement to the class as a whole, but individuals are also singled out for meritorious contributions and awards (or bonus point give aways) are announced.

The Expanding Superdocument

By the end of the semester, there may be as many as 100 links in any student's Web document, connecting themselves to others through some idea expressed in a specific paragraph. In one semester of 16 weeks, two interlinked classes of 20 students each, typically produce a total of 12 megabytes of text containing about 8,000 links. Each semester a new crop of students peruse the previous generation documents and write their own documents interlinked with each other and with the prior semesters. The interlinked reports within and across generations constitutes an ever expanding virtual superdocument. Figure 1 presents a conceptual representation of this hypertext document.

New students form a fresh cohort each semester. The first generation leaves behind its interlinked reports. The second generation works itself through the existing linked documents and creates their own. The third generation goes through the first and second, and leaves its own. The hypertext superdocument keeps expanding as each generation processes the previous one and adds its layer to the top. Throughout this community-building process the instructor's weekly homework assignments play a central directive function. Reading, discussing, annotating, and linking are the student activities that the instructor needs to guide. Feedback , evaluation and correction are essential for learning to continue. Keeping up with this task has required a major investment of time on the part of the instructor. However, this kind of repeated close inspection of student work and contact with their ideas has been for the instructor a source of surprise, delight and deep satisfaction.

The Virtual Size of a Hypertext Superdocument

Is a hypertext superdocument real? Can it be identified through the physical, URL locations of a collection of files on some Web server's computer disk? Of course, the physical files are a pre-requisite condition for the superdocument to exist, but the physical files do not define the hypertext superdocument. Rather, one can say that its existence is virtual, in the sense that its hypertext links create a cyberspace highway whose existence is beyond the physical files and cannot be defined by them. The World Wide Web as a whole, defined as thousands of interlinked Home Pages across the Internet, is itself a very large virtual superdocument that can be traversed in a different way by millions of cybernaut browsers in an endlessly variable patterns. The number of possible ways of traveling the Internet is endless and its size may be called virtual infinity. Each browsing session has the potential of uncovering new information, new connection of ideas, new ways of linking concepts.

The more a link is traveled the greater its value as an information resource. The more links exist in a hypertext document, the richer is the document's cultural value. How can one calculate the virtual (rather than actual) size of a hypertext superdocument? One approach might be to treat links as a multiplicative factor functioning to magnify the potential traffic on the information superhighway. For instance, the first generation of 40 students in this cyberspace learning community produced 12 megabytes of disk space and created a total of 4,000 links. Multiplying these two figures we obtain the value of 48,000 megabyte-links or megalinks as the size of the cyberspace created by Generation 1. This is an incomplete formulation since it assumes that all links are equivalent, which in fact they are not, as will be now shown.


Three categories of links are observable in the superdocument: within each generation, across generations, and links to the outside world. Further, the links within each category there may be either two-way or one-way. Two-way links are mutual or reciprocal. That is, a browser may see a paragraph in a file belonging to student A, with a link to student B. Traversing that link brings the browser to student B's paragraph where a link is found to the paragraph in student A. The mutuality of two-way links enhances the information value of links. Two-way links are possible across generations if there is a mechanism in place for inserting links backwards in time, that is, from generation 1 to generation 2, and so on. These links may be called anachronistic since they allow a cybernaut who is browsing in generation 1 or 2 to take jumps to generations 3 or 4, and so on.

Anachronistic links are created by the instructor as part of the maintenance requirements of the generational superdocument. This is a labor intensive effort that requires regular scanning of the current student files to identify one-way links that may have been created to prior generations, and then creating reciprocal anachronistic links in the older documents to the new. Without the provision for creating anachronistic links, the older parts of the superdocument would gradually die and become an archive or database. These have their proper function, and several types of archived student files have been included in the overall Web documents available to the students. It is necessary to create anachronistic links in order to maintain the virtual superdocument as a live cultural resource for knowledge creation in a generational cybercommunity.

In our educational experiment so far, we note the occurrence of seven types of links in this cyberspace learning community:
Within generations:

1) Two-way links between two students or a student and the instructor
2)One-way links from one student to another student or the instructor
Across generations:
3)One-way links (created by students) from student X in one generation to student Y in some prior generation
4)Two-way anachronistic links between student X in one generation to student Y in some prior generation
5)One-way anachronistic links (created by the instructor) from student M in one generation to student N in some subsequent generation
To outside world:
6)One-way link from a member of the cyberspace learning community to the outside World Wide Web
7)Two-way links between a member of the learning community and the outside world (this requires outside world cooperation).
An additional possibility exists, namely one-way links from the outside world to a student in the cyberspace learning community. However this would be difficult to locate, given the enormous size of the World Wide Web cyberspace.


Maintenance requirements of the hypertext superdocument
As the number of generations grows the virtual superdocument grows at a certain rate which is either fixed or variable. Figure 1 currently depicts a model that assumes a constant amount of increment with each generation. This means that as the superdocument grows in megalinks size the cyberspace thins out, in the sense of less links per unit of text. This is because the number of links created by each generation is limited by the semester hours available. For a three-credit college course, we can assume an average limit of around 100 links per student per semester with approximately 30 typed pages of contributed text. As the size of the text grows generationally more and more links would have to be created by each new generation in order to maintain the same density. But this is not a realistic expectation. Hence the density of the cyberspace will tend to thin out. Continuing observation of the growing superdocument will reveal how it grows in terms of size, density, and shape. Methods need to be discovered for ways of maintaining the superdocument up to date and of creating anachronistic two-way links where appropriate.

Psychological properties of hypertext links in a cyberspace learning community

Analysis of links must include the question of what motivates a member of a cyberspace learning community to create a particular link. This may require content analysis of the text containing a link as well as the text to which the link refers. Additionally, members may be interviewed or prompted by a questionnaire requesting information on the motive for creating a particular link. So far five types of motives have been observed for creating links:
1) Wanting to facilitate browsing or navigation
2) Providing a topical index for identifying live issues and for easier access to them
3) Wanting to entertain browsers by providing access to curiosities or fun things to see or do
4) Seeking confirmation or validation for one's opinions or feelings by linking to another person's comments or acts
5) Illustrating or proving one's point by linking to other data or information.

Natural vs. Artificial Superdocuments

The growth of a virtual superdocument within a generational cybercommunity can be viewed and studied as an anthropological and ethnolinguistic phenomenon. Its study and analysis over time leads to new knowledge about community-building forces that can be useful to cyberspace managers who are participants and leaders in cybercommunities such as system administrators, discussion group managers, Home Page architects, electronic mall owners, Internet trainers, and instructors of cyberspace learning communities. However, the validity of data obtained from the analysis of a superdocument may depend on the authenticity of the cybercommunity that produced it.

A natural superdocument is the virtual outcome of the social and psychological exchanges transacted by members of a cybercommunity as part of their normal and regular membership activity over time. The virtual superdocument discussed in Fig. 1 is being produced by generations of college students, with semester-cohorts of 40 students, and 16 weeks of activity per semester, and a real environment made up of course credits, homework assignments, and class discussions. The assignments involve both the production of scholarly text ("weekly lab reports on your Web document") and the critical annotation of the work of other students ("comment on their work and make links to their paragraph"). The text they write and the links they create are thus valid data for anthropologists and students of human behavior.

An artificial superdocument, on the other hand, is produced by one or more indexers or publishers who create interconnected text collections on the basis of their subject content using keywords and cross-references to facilitate searching, navigating, teaching, vending, entertaining, and so on. The virtual universe of gopher directories on the Internet is an instance of an artificial superdocument. Another artificial superdocument is the interlinked network of WWW indexes and directories ("Other places to visit" or "My favorite bookmarks" etc.). No doubt future research will study the differences between artificial and natural superdocuments.


[1] J. Conklin, "Hypertext: An Introduction and Survey," Computer, vol 20, pp. 17-41, Sept. 1987.

[2] N. Yankelovitch, B. Haan, N. Meyerowitz, and S. Drucker, "Intermedia: The Concept and the Construction of a Seamless Information Environment," IEEE Computer, vol. 21, pp. 81-96, Jan. 1988.

[3] M. Mohageg, "The Influence of Hypertext Linking Structures on the Efficiency of Information Retrieval," Human Factors, vol. 34, pp. 351-367, June, 1992.

[4] F. Halasz, "Reflections on NoteCards: Seven Issues for the Next Generation," Communications of the ACM, vol. 31, pp. 836-852, July, 1988.

[5] W. Horton, Designing and Writing Online Documentation, New York: John Wiley and Sons, 1990.

[6] H. Rheingold, The Virtual Community, New York: HarperPerennial, 1994.

Click here for Picture

Figure 1. The expanding generational virtual hypertext superdocument in a cybersapce learning community.

Are you a visitor? I would appreciate your reactions or comments. How did you get here? What did you see?

Leave a massage for Leon James?

Back to the top