The 3rd workshop on hypertext functionality was held at the ACM Hypertext conference held in Southampton, England, April 1997. As an introduction to the fourth in this series of workshops, we present a summary of the previous session. First, we present a short summary of the results of the workshop and then the overview of the position the HTF working group to the Open Hypertext working group.
HTF3 Summary
Hypertext is not necessarily the best approach for every system, rather its effectiveness depends on the domain and task. Hypertext seems to be appropriate when one or more of the following conditions exists:
Any methodology for hypertext design includes three steps: definition of task domain, definition of hypertext elements, and definition of relationships. Hypertext is not meant to replace software engineering tools, for example, but to facilitate the or ganization of thoughts and data to a level that lower level tools can be used effectively.
The issues of hypertext functionality are not based on the notion of intra vs. extra. Rather we propose that functionalities could be classified as task independent and task specific. For example, link management is a generic task independent function while customized link services could be generated for specific tasks.
Definitions of nodes and links are too restrictive. In particular nodes may be processes as well as data and links may include hand-crafted links, private vs. shared links, adaptive links, and temporal links (i.e.,based on successive accesses to data (i.e. first...nth))
Although many of the problems of distributed hypertext systems are largely ones of operating system level (such as locking and access mechanisms) rather than information system level the following issues are relevant to hypertext: trader vs. broker, le vels of trust and co-operation, management of cross organizational boundaries, and management of privileges. Locking mechanisms and access controls are required for management of collaborative and shared data sets.
Navigation in the temporal dimension requires version control, notification of updates, and consistency for both synchronous and asynchronous usages. The use of Bayesian networks and fuzzy logic for modeling and using uncertainty in design and navigati on of hypertext systems. BN's include mechanisms for dynamic updates and are proposed as support for opinions and subjective knowledge in particular.
The HTF working group takes a task/user orientation to the definition of hypertext functionality. First we need to determine those task domains for which hypertext is the most appropriate approach. Second, we need to define the high-level, generic hype rtext functionalities that are necessary to support these tasks. Third, we need to develop benchmarks and metrics for evaluating hypertext systems.
We define hypertext as a mapping mechanism from data to the user's current requirement. That is, hypertext provides support for user-centric views of data. We have established that the hypertext approach (i.e., inclusion of hypertext functionalities) i s indicated were any of the following conditions exist:
We need now to define high-level functions that we would require in a hypertext toolbox to support hypertext in these circumstances. What functions are required to support user-centric views on data-centric structures. We propose the development of sev eral task-oriented scenarios as a first step in establishing such a set of generic functional requirements; design rational support in software engineering, www search process, and collaborative learning module. The scenarios will be developed in use-case and/or OHS format for scenario design. The functional requirements from this high-level analysis will form the basis for discussion with the OHS.
Benchmarks of performance and effectiveness are necessary to show the effect of hypertext functionalities in systems.
More Details
For more details on HTF3 you can access the full report and position papers.