aJodrey School of Computer Science, Acadia University, Canada
bFaculty of Computer Science, Dalhousie University, Canada
The protocols and browsers available on the Web provide robust hypertext-style access to data stored worldwide. Why then, are not all hypertext applications built on the Web? In general, "hypertext functionality" goes beyond browsing by clicking on "goto" links from one document to another. In this short paper, we use, by way of example, two hypertext applications that have been implemented on the Web; an interactive mathematics dictionary and the Digital Agora. We then suggest improvements in a relatively few areas that would result in a Web environment much more hospitable for hypertext applications.
Interactive math dictionary — The MathResource [5,3] was originally a standalone electronic version of a math dictionary of over 4000 mathematical terms. The MathResource differs from other electronic dictionaries by offering interaction with the mathematical content as well as regular hypertext links. Users can link from expressions in the text to a computational environment to alter and execute expressions generating numeric results and/or plots. Several difficulties arose in the deployment of the MathResource in a Web version. First, two distinct link types are required: hypertext links and computational links. These links must be distinguishable to both the user and the system. They need to look different as well as act different. Second, support is required for the sophisticated display of mathematical expressions and plots that are generated dynamically.
Digital Agora — The Digital Agora [4], Fig. 1, is a hypertext application built on the Web to support collaborative learning in political science. The system provides hypertext tools to facilitate understanding of complex issues, such as peace initiatives or bioethics, from multiple points of view. The students, typically in groups, access information, formulate analysis of the problem, and communicate results to others. We decided to implement this system on the Web to take advantage of existing facilities for search, communication, and presentation. The Web version falls short of expectations in several areas. First, the process of argumentation and consensus building requires links that are semantically useful, such as in support of, example, big picture, question, annotation. Second, collaborative work requires the notion of links that belong to members of a group and are invisible to others. Third, students need to be able to author links from documents that they do not own, i.e., third-party documents, so that these documents are seen as part of the analysis. Fourth, the roles of reader and author need to be merged for the students to develop analyses collaboratively. Finally, semantically enriched visualizations are required. An important component of critical thinking that the students employ while formulating understandings of complex issues is the creation of lateral maps [2], Fig. 2. As long as these concept-style maps are treated as bit maps, they are not interactive, can not be searched, linked, or worked on in collaboration effectively.
Based on our experiences, we suggest that improvements in a relatively few areas would result in a Web environment that would be much more hospitable for these types of hypertext applications.
Links — The simple "goto" link supported by the Web is too impoverished to support many of the new hypertext applications. Suggestions abound [1] for improvements in link structures but we emphasize two: computational and collaborative.
Computational links — From the MathResource project we see that not all links are reference links but that a class of links is needed that supports the computation of results or the invocation of processes that generate results that the user may need.
Collaborative links — From the Digital Agora project we see that classes of links are required that support collaborative activity. In addition to semantically typed links to support argumentation and point of view, links to annotation and links from third-party sites for discussion support are very important. Once links have semantics and labels, such as those in Fig. 3, they can be searched and clustered to support collaborative tasks.
Authoring — Each of these projects promotes the merger of reader and author roles of the users. Two way interaction is required and expected. Users need to be able to initiate computation, annotation and semantic links of the types described above while reading.
In particular, students in collaborative groups have a combined reader/author role, with little if any distinction between the functions.
Version control — The Digital Agora requires systematic version control of information and of interactions. The temporal dimension of the data and of the links in these cases has semantic value for the users.
Graphics — Currently, the Web treats all images as bit-maps. In the Digital Agora lateral maps, which are basic elements of discourse, are empoverished as bit maps. Support is required to preserve the semantics of components of visual presentations just as the semantics of textual presentations is maintained. In the MathResource, plots must remain dynamic objects that the user can manipulate rather than static images.
Access levels — Particularly in the Digital Agora some transparent implementation and management of access levels is needed. At times groups need to share ideas and relationships within the group until consensus is reached. At other times the presentation may be made accessible to larger audiences. Students, of course, may be members of multiple groups at any time.
The agenda for future Web development presented here is based on our experience with implementation of hypertext applications on the Web. We are not presenting this as an exhaustive list of future development requirements but rather as a basic set of improvements that will enable the Web to support a wide range of hypertext applications and for hypertext applications to exploit the potential of the Web.
[1] M. Bieber, F. Vitali, H. Ashman, V. Balasubramanian, and H. Oinas-Kukkonen, Fourth generation hypermedia: some missing links for the World Wide Web, in: International Journal of Human-Computer Studies, 47(1), 1997, http://ijhcs.open.ac.uk/bieber/bieber.html
[2] E. DeBono, Lateral Thinking for Management, McGraw-Hill, New York, NY, 1971.
[3] MathResources, [http://www.math.mathresources.ca]
[4] C. Watters, M. Conley, and C. Alexander, The Digital Agora: using technology for learning in the social sciences. Comm. of ACM, January 1998, pp. 50–57.
[5] C.R. Watters and J. Ho., MathProbe: active mathematical dictionary, in: RIAO 94 Conference Proceedings, New York, October 11–13, 1994, pp. 552–569.
Fig. 1. Digital Agora home page.
Fig. 2. Lateral map.
Fig. 3. Typed links.
