Cybermountain logo

<--Word Circuits ---->Information about CyberMountain Colloquium and the corresponding MOO

Realizing the potential of hypermedia as a collaborative artistic medium

Peter J. Nürnberg
Department of Computer Science, Aarhus University

Abstract. Hypermedia, from its inception, has always been conceptualized as a medium for supporting (both synchronous and asynchronous) collaborative work. Theoreticians have often linked hypermedia's polyvocal attributes with a kind of implicit asynchronous dialogue among "wreaders". However, artists trying to explore the collaborative aspects of hypermedia have been hindered by systems that offer hypermedia functionality only within single user (or highly asynchronous) environments. This paper explores some of the reasons why collaboration support has not been more generally available to hyper-artists, and current ongoing work to correct this situation.


Even the very earliest hypermedia researchers (e.g., Bush [1945], Engelbart [1962], Nelson [1967], etc.) called for hypermedia to be used to support collaborative work. With its ability to support multiple structures over the same set of data, hypermedia allows a realization of polyvalence unlike those available in other traditional media. Others have intimately linked hypermedia functionality with providing readers the ability to annotate work owned by themselves or others, thereby explicitly blurring the distinction made by theoreticians between reader and writer. As useful as hypermedia may be for individual tasks of associative storage and recall, it seems to be a medium, the potential of which can only be fully appreciated in collaborative settings.

Despite this, collaboration support in hypermedia systems has generally been a second-class feature, hastily added (if at all) to hypermedia systems without appreciation for the specific requirements generated by the special features of the hypermedia itself. Rarely can one find a system that supports true synchronous work; more rarely still one in which distinctions between classes of users (e.g., authors versus annotators versus browsers) can be made fluid or made even to dissappear - a necessary precondition to support of collaborative work in real-world, dynamic situations. Furthermore, systems designed to support hypermedia writers (or, more generally, any type of hypermedia artist) seem especially insufficient with respect to collaboration support, despite the strong theoretical foundations for developing such functionality.

Hypermedia System Research

Collaboration support in hypermedia systems has been slow incoming largely because of a "split" in the hypermedia community. This split exists between "infrastructure" systems people and "domain" systems people.

Infrastructure people have worked on increasingly complex and flexible infrastructures that have until recently only been applied to extremely limited domains and required specialized hardware, software, or operating systems. Infrastructure work has tended to focus on the "backends" of systems, with less thought being given to how hypermedia is presented. For infrastructure people, the definition of hypermedia at the frontend has remained relatively constant since the time of Bush - "associative storage and recall" hypermedia, in which people link nodes of information together, sometimes constructing trails of such associations. The focus of infrastructure work is to do such work "better" (i.e., faster, more efficiently, over a wider scale, etc.)

Domains people have worked on hypermedia systems designed to address an ever widening array of problems. They have applied hypermedia structuring concepts to information analysis (spatial hypertext), argumentation support (issue-based hypertext), classification work (taxonomic hypertext), and other areas. Domains people have added whole new sets of structural abstractions beyond node and link - spaces, relations, taxa, composites, etc. As such, these systems tend to focus on the frontend, concentrating on how people interact with such new abstractions. Less attention has been paid to building flexible, "real-world" backends that scale well, allow multiple users to work collaboratively, etc.

Within the last three years, the Open Hypermedia Systems Working Group (OHSWG, see [1997]) has focused on bringing to bear the developments of the last two decades of infrastructure work to the wide variety of problems the domains people have addressed in that time. Some initial successes have been demonstrated at the last two ACM Hypertext conferences (1998, Pittsburgh, and 1999, Darmstadt). The OHSWG demonstrated collaborative hypermedia to handle "standard" node-link structures, spatial hypertext (CAOS, see [Reinert et al. 1999]) and content-based retrieval (Solent, see [Reich et al. 1999]). These new systems generally fall under the rubric of "component-based open hypermedia systems" (CB-OHS's).

The two main design principles for CB-OHS's so far have been openness and integration. Openness stresses the ability to integrate new applications and structure services into the environment with the least amount of effort possible. Applications should be able to be integrated to use structure services orthogonal to other services (such as persistent storage). Integration stresses the ability of different structure services to be able to work together, sharing structures and data, to allow applications to view the same structure different at different times, to suit the chaging needs of users.

Significance For Writers

Such new collaborative, open, and well-integrated CB-OHS environments offer many new possibilities for writers and readers of hyperfiction. Currently, writers and readers are constrained by the expectations of the authors of the software they use, which no matter how broad and well-intentioned, never take into account all possible future use situations. The call for extensibility has been made many times by both system users and developers. CB-OHS's offer a realization of this design goal. Experiments are currently underway to integrate the Storyspace system into the Construct CB-OHS [Wiil and Nürnberg 1998]. We look forward to reporting on the results of these experiments in the near future.


[Bush 1945] Bush, V. 1945. As we may think. Atlantic Monthly, 176(1), 101-108.

[Engelbart 1962] Engelbart, D. E. 1962. Augmenting human intellect: a conceptual framework. Tech. Report AFOSR-3223, Contract AF 49(638)-1024, Stanford Research Institute.

[Nelson 1967] Nelson, T. H. 1967. Getting it out of our system. Information Retrieval: A Critical View, G. Shector, Ed. Thompson Book Co., Washington, DC, 191-210.

[OHSWG 1997] <>

[Reich et al. 1999] Reich, S., Wiil, U. K., Nürnberg, P. J., Grønbæk, K., Davis, H. C., Anderson, K. M., Millard, D., and Haake, J. M. 1999. Open hypermedia protocol. submitted to New Review of Hypermedia and Multimedia special issue on OHS.

[Reinert et al. 1999] Reinert, O., Bucka-Lassen, D., Pedersen, C., and Nürnberg, P. J. 1999. CAOS: A collaborative and open spatial structure service component with incremental spatial parsing. Proceedings of ACM HT '99.

[Wiil and Nürnberg 1998] Wiil, U. K., and Nürnberg, P. J. 1999. Evolving hypermedia middleware services: lessons and observations. Proceedings of SAC '99, (San Antonio, TX, Feb), ACM Press.