The Human Element in Collaborative Hypertext, by Murray Turoff, Starr Roxanne Hiltz and V. Balasubramanina. Short Workshop paper presented at the workshop on Collaborative Hypertext held at the CSCW 1994 meeting.

The Human Element in Collaborative Hypertext\Hypermedia

by

Murray Turoff, Starr Roxanne Hiltz & V. Balasubramanian

New Jersey Institute of Technology

University Heights, Newark NJ, 07102

201 596 3399

murray@eies.njit.edu

When computers were introduced in libraries, the very first thing they accomplished was to destroy beneficial forms of indirect communications. For those of you who can remember, there was a day when the library card had the signature of everyone who had borrowed the material whether it be a book, document, report, or record. With this mechanism, one was able to discover who within the local community had the same interests, tastes, or concerns as yourself. This was a form of indirect communication channel (Turoff, 1991) that could be extremely beneficial. In research libraries it was invaluable knowledge.

The processes of reading and communication are inseparably linked. What we read not only enriches our ability to communicate; reading is but a part of a greater communication process. This should be a basic assumption in the design of Hypertext (which we use in the original meaning of the word "text" that included graphics, pictures, etc.). This leads to the first of the propositions or premises which we wish to present at the workshop about fundamental principles which should be followed in the design of collaborative Hypertext systems:

P1. Hypertext systems need to be viewed within the more holistic context of communications: between "authors" and viewers, among co-authors, and among users.

P2. In collaborative Hypertext/Hypermedia any user should also be able to be an "author" at least in the sense of indirect communication of information to others.

An example of an indirect communication of information might be an automatic increment in counts of how often a node or link has been: "visited," "utilized," or "appended." Over time such counts might be useful in both providing guidance and in restructuring the database. In fact, knowing what has not been utilized could be a useful indicator of a poor Hypertext structure within the database. Also knowing what users are looking for and not finding is tremendously useful for an evolving database (Hiltz & Turoff, 1993).

It is also of interest to note the very large number of ways in which people go about marking up already created material. In a college library many students prefer to get a copy of a paper so that they can mark it up. Most used text books in the bookstore have extensive markups in them. People use special markup pens in different colors to indicate different types of information important to them. Readers create index cards with notes, draw lines between words or paragraphs to indicate relationships not obvious in the text, and move back and forth in the text rather than following the original linear path. Readers are attempting to organize and record a wide number of reactions to the material, such as:

* Forming degrees of agreement or disagreement with the material

* Seeking or reaching an understanding of the material

* Relating it to what one already knows (or doesn't know)

* Realizing confusions and lack of understandings that need further pursuit

The processes of reading and writing are also inseparably linked. It is an over-simplification to design systems that separate the reading functionality from the composition functionality. Efforts to treat these two areas as different problems are not consistent with designing systems for humans.

P3. Effective Hypertext systems must integrate the functionality of reading and composition.

Those of us who are educators know that a significant part of the knowledge we attempt to convey to learners is the problem conceptualization and problem solving cognitive processes that have evolved as appropriate to that field of study. We are teaching a way of thinking, a thought process. Even in a technical course, it is not the finished solution to an analysis problem or the expression of a derived formula that is important. Rather, it is the thinking process that allows a person to derive a result or solve a problem, and the collection of individual steps or the expression of the thinking process. One does not learn to paint by looking at finished painting but by observing a good artist execute the actual process of creating the painting. The design of a product, a building, a bridge, a software system all involve specialized design processes that the experts have acquired from years of experience.

A generally unrecognized part of the problem of "becoming lost in the web" of Hypertext is that most systems deal with capturing a finished result and not with capturing the process whereby the results were created in the first place. Different people with different backgrounds and viewpoints will design very different results if given the same problem. From the Delphi area (Linstone & Turoff, 1975) we know that when dealing with heterogeneous groups of experts, that collaborative understandings can only be evolved when the group is able to view the content of the process that led to different results among the members of the group. Many Delphi designs were potential process oriented Hypertext Systems.

P4. Effective Hypertext systems must provide for the recording and animation of the creation and modification process.

If a Hypertext system is a knowledge base, then the path that different people take in exploring vast stores of knowledge can be a representation of their cognitive problem solving process in dealing with that topic. A student following the path of a teacher is exposed to the expert's view. A teacher following the path of a student gains an understanding of what the student has learned. The view of a political scientist and that of engineer on what is important in a database governing a particular environmental issue in a community can be very different. However, each view can be significant for those trying to make an evaluation.

P5. Effective Hypertext systems must provide for recording and playback of the reading and associated modification process.

One cannot help but mention that knowing how the different candidates for a public office go about viewing a Hypertext version of the New York Times might be the most informative way on deciding whom to vote for. Similarly it is clear that capturing the thought processes of a human trying to explore a complex knowledge base provides an extremely informative insight into their thinking process. Such functionality has to be offered with informed consent on the part of the users, of course.

When one considers that in our experiences with Computer Mediated Communications (Hiltz & Turoff, 1993) even small groups of people working together on a complex problem can generate thousands of comments in the space of a few months. In CMC information overload by collaborative groups has always been a fundamental problem [Hiltz & Turoff, 1985]. There is not any real difference between the problems of a collaborative Hypertext and a CMC system. Information retrieval and indexing are key to the solution of this problem. In addition, it is the mediation among the members of the group that provides the very necessary foundation for approaching this problem. Groups that work together on a regular basis convert ambiguity in terms and concepts to very specific meanings, and invent of new terms and acronyms.

P6. Groups must be able to imposed a shared view and their own evolving collaborative indexing approaches on a collaborative Hypertext system.

People are not going to use separate Hypertext systems tailored to specific problem areas. Humans are already overloaded with many different packages they must learn to do different things. This means that there is a real need to evolve a true general purpose Hypertext system that can be applied to any problem. One approach is to make the underlying data structure independent of semantics. However, this leads to high levels of system opacity between the users and the system. Another approach is that the database structure must allow for comprehensive general purpose semantic meanings on both the links and the nodes. We feel that the best way this can be done is to utilize a structure that reflects general purpose human intellectual ability to formulate concepts and relationships (Turoff, et. el, 1991; Rao & Turoff, 1990). Our particular approach is based upon an existing model of the human intellect from cognitive psychology:

Hypertext Morphology

(Based upon Guilford's theory of the human intellect)

Guilford:CognitionConvergent
Production
Divergent
Production
ProductHypertext:
NodesConvergent LinksDivergent Links
unitsdetailspecificationelaboration
classescollectionmembershipopposition
relationspropositionassociationspeculation
systemssummarypathbranch
transformationsissuealternativelateral
implicationsobservationinferenceextrapolation

This approach is parsimoniously beautiful in that it says all forms of human cognition's can be expressed by six node types and twelve link types. However, it is not the particular model that is important, since much research would have to be done to determine what is the best model. What is important for those working in the Hypertext field is a recognition that:

P7. Effective collaborative Hypertext systems must impose general purpose semantic structures upon Hypertext databases, that allow a wide variety of applications and thought processes to be represented.

Note that capturing the thought process in the above classification becomes a simplified matter of capturing the creation steps in time order.

References:

[Hiltz & Turoff, 1993]. Hiltz, Starr Roxanne and Murray Turoff. The Network Nation: Human Communication via Computer, Revised Edition, MIT Press, Cambridge, Mass. 1993.

[Hiltz & Turoff, 1985]. Hiltz, Starr Roxanne and Turoff, Murray, (1985), Structuring Computer-Mediated Communication Systems to Avoid Information Overload. Communications of the ACM, 28, 7 (July): 680-689.

[Linstone & Turoff, 1975]. Linstone, Harold, and Murray Turoff, eds. The Delphi Method: Techniques and Applications, Addison Wesley, Reading, Mass. 1975.

[Rao & Turoff, 1990]. Rao, Usha and Turoff, Murray, Hypertext Functionality: A Theoretical Framework, International Journal of Human Computer Interaction, 2(4), 1990, pp. 338-358.

[Turoff, 1991]. Turoff, Murray. Computer-Mediated Communication Requirements for Group Support, Journal of Organizational Computing, 1(1), 85-113, 1991.

[Turoff, et. al., 1993]. Turoff, Murray, S. R. Hiltz, A. N. F. Bahgat, and Ajaz Rana. Distributed Group Support Systems, MIS Quarterly; December 1993, 399-417.

[Turoff et al., 1991]. Turoff, Murray, Rao, Usha, and Hiltz, Starr Roxanne. Collaborative Hypertext in Computer Mediated Communications, Proceedings of the twenty-fourth Annual Hawaii International Conference on System Sciences, January 1991.

Note on author qualifications: Murray Turoff and Starr Roxanne Hiltz are Distinguished Professors of Computer and Information Science at New Jersey Institute of Technology. They have been working with collaborative systems for about 20 years, and are currently in the process of equiping a new Collaborative Hypermedia laboratory, in preparation for a new series of research and development projects in this area. V. Balasubramanian is one of several Ph.D. candidates currently doing dissertation research in this area, under the guidance of Turoff and Hiltz.


Back to homepage of Murray Turoff
Back to homepage of Roxanne Hiltz